Interview guides · Energy Transfer

Energy Transfer Software Engineer Interview Questions & Guide 2026

Debugging practice for Energy Transfer software engineer candidates: reproduce it, narrow it, then change one thing per run.

The 60-second brief

A Software Engineer at Energy Transfer operates at the intersection of critical infrastructure and cutting-edge technology. Whether you are working on Advanced Process Controls, Pipeline Integrity systems, or AI development, your work directly impacts the safety, efficiency, and reliability of one of the nation’s largest energy networks. This role is not just about writing code; it is about building robust, scalable solutions that manage complex real-world data in high-stakes environments.

The title spans product, platform and infrastructure work, and which of those the seat actually is decides whether design or algorithms carries more weight in your preparation. The posting rarely settles it; what the team is on call for usually does.

Focus areas

Keep a site operating during a network partition
Reconcile a movement ledger against derived positions
Fold late, out-of-order scan events by event time

Overview

You will contribute to projects that solve unique challenges, such as optimizing energy transport flows or leveraging artificial intelligence to enhance operational decision-making. The scale of the work is significant, often requiring a deep understanding of how software integrates with hardware and industrial systems. For a Software Engineer here, the impact is tangible: your software ensures that energy resources move safely and effectively across the country.

Candidates should emphasize their ability to translate complex technical requirements into reliable, production-ready code that meets rigorous industrial standards.

The Energy Transfer interview process

Energy Transfer candidates report 3 rounds · ≈ 3-5 weeks. The stages below are what candidates describe, not a published process.

Screening Round

When a round has no standard shape, it is often there because something is still open: an area no earlier conversation reached, a round where the signal came out mixed, or a decision someone is not ready to make alone. Work out which by going back over what each earlier round actually covered rather than how it felt, and arrive able to give evidence on that point without being asked twice. Weak answers replay the loop's earlier material at the same depth. Strong ones go a level deeper and stay consistent with what you already said.

What to demonstrate

Whether your account of a project matches the one you gave earlier in the loop, since what you said before may be available to whoever runs this round
Whether you can go a level deeper on something already covered, reaching the decision and its alternatives rather than repeating the summary
Whether you state your own uncertainty accurately, including parts of a system you did not build and decisions you inherited, instead of claiming even ownership across all of it
Whether you can answer a question you handled poorly earlier by naming what you missed, rather than delivering a polished second version as if the first had not happened

How to prepare

Reconstruct the loop on one page: for each round, the questions you were asked and the answer you actually gave, not the better one you thought of afterwards. The gaps on that page are your best available guess at why this round exists.
Take the two claims you made earlier that carry the most weight and assemble the backing for each: the measurement, the date, what broke, the decision you would make differently now.
Write down the three facts about your work that must not drift between tellings, such as team size, timeline and your own role, and check your stories against that list rather than trusting recall under pressure

Technical Interviews

The same problem is scored by two different mechanisms depending on the format, and preparing for one does not cover the other. With a person watching, partial progress is visible and a hint is a correction you can absorb; silence is the expensive failure, because nobody can read a half-written function. With an automated grader there is no partial credit for what you were about to do, nobody to ask, and the worked examples in the prompt are the entire specification. Read them as a contract, down to whether an empty result should be an empty list or no output at all.

What to demonstrate

In a live session, whether your commentary tracks what your hands are doing, and whether a hint redirects you or gets defended against
In an automated one, whether you cover the cases the examples do not show, since the hidden cases are where the score moves
Whether you manage the clock on purpose: abandoning an approach that is not converging while there is still time to write something simpler that finishes

How to prepare

Have someone hand you a problem and feed you one deliberately wrong hint. Practise testing it against a concrete case instead of accepting or rejecting it on authority.
Do one timed run a week in a plain browser editor with autocomplete, linting and your own snippets switched off, which is closer to what these environments give you
For the automated format, write the harness before the solution: a main that feeds the worked examples plus an empty and a single-element case and prints expected against actual, so a wrong submission is caught by you first

Behavioral Interviews

Your first answer is not really what is scored. It buys the follow-up questions, and those decide the round. An interviewer with fifteen minutes takes one thread and pushes on it four or five times, so a story you can only tell at a single level of detail collapses under the third why. That is an argument for fewer stories known deeply rather than one prepared per prompt. Four or five pieces of work you can still explain down to the code you changed and the argument you had about it will cover nearly anything asked in this round.

What to demonstrate

Whether a story holds as the questioning moves from what you did to why that instead of the alternative, and then to what you would change knowing what you know now
Whether you can re-cut a project to answer the question actually asked rather than delivering a rehearsed block that answers an adjacent one
Whether your level of detail is chosen rather than habitual: going down to the schema when the question is about the data model, staying out of it when the question is about the person who disagreed with you

How to prepare

Pick four projects and write the chain out four levels deep for each: what you did, why that, why not the alternative, and what would have to be true for the alternative to have won. Where you cannot reach the fourth level, you have a placeholder rather than a story
Have someone ask why three times in a row on a single thread with nothing else added, and mark the point where you start repeating a sentence you already said. That point is where the interviewer stops learning anything
Build a one-page index instead of an answer bank: the common prompts in this round (disagreement, a failure that was yours, thin requirements, a deadline you missed, work you inherited) mapped to which of your four projects you would use for each, so the choosing is done now rather than while an interviewer waits

Where candidates lose points

Treating the database as the truth about physical stock

Every system in this domain drifts from reality, through miscounts, unrecorded damage, mis-scans and theft, so a design with no adjustment path forces the drift somewhere worse. The usual shape is a CHECK constraint or application guard that clamps a position at zero: the negative that would have exposed a missing receipt is silently swallowed, and the first visible symptom is a picker standing at an empty location for an order the system promised. The correct posture is to let the invariant fail loudly - record the cycle-count adjustment as a first-class movement with its own reason code, surface the variance, and let planning see that this location is unreliable - because an adjustment that is indistinguishable from a normal pick destroys the only signal that would have found the root cause. A related version of the same mistake is deleting or updating a movement row to fix a mistake, which leaves the position correct and the history a lie.

Holding a database transaction open across a physical or third-party operation

It is natural to open a transaction, lock the position, call the rating or tendering API, and commit on the response, and it works perfectly until the partner's p99 goes from 200 milliseconds to thirty seconds. At that point every request holding a lock on a hot item-node pair queues behind it, the connection pool fills with transactions that are waiting on the network rather than on the database, and an unrelated service sharing the pool fails at the same moment. In this domain the effect is amplified because demand concentrates on a few hot rows during a promotion or a seasonal peak, exactly when partner latency is also degraded. The structural fix is to keep transactions short and local - commit the state change together with an outbox row, let a relay perform the external call, and reconcile asynchronously - accepting at-least-once delivery and making the effect idempotent rather than trying to stretch a transaction over something the database cannot roll back.

Writing code before the input contract is pinned down

Before the first line, state the types, the size bounds, whether duplicates, negatives or an empty input are possible, whether the input is sorted, whether you may mutate it, and what the function returns when nothing matches. Every one of those answers changes the code, and discovering one at minute twenty costs a rewrite you no longer have time for.

Comparing floating-point values for equality, or holding money in them

Binary floating point cannot represent 0.1 exactly, so repeated addition drifts and an equality check fails on values that are mathematically equal. Store currency as integer minor units or a decimal type, and compare floats against a tolerance you chose for a stated reason.

Questions & practice

Find the hottest write keys under a fixed memory budget

Medium · top-k · heap · streaming sketches

One day of writes to inventory_movement is 200 million rows; each carries (item_id, node_id). Distinct pairs reach 60 million. Return the 200 pairs with the most writes so they can be moved behind a single writer before the next peak. First give the exact answer assuming you may hold 60 million counters, then give an answer under a 200 MB budget, and state the error guarantee that second answer carries. Both must be one pass over the day's rows unless you argue for a second.

Approach

Exact version: one hash pass to build (item_id, node_id) -> int64, then a single scan of the map maintaining a min-heap of size K = 200, pushing and popping when the incoming count exceeds the heap root. O(n + d log K) with n = 2e8 and d = 6e7, versus O(d log d) if you sort all 60 million entries — a factor of roughly log d / log K, about 3.4x, for no benefit. The heap must be a min-heap: the root is the weakest survivor, so eviction is O(log K).
Cost the map honestly. 60 million entries at a 16-byte key, an int64 count and per-entry overhead is 2-3 GB in most runtimes, which is the whole reason for the second version rather than a theoretical aside.
Bounded version: Misra-Gries with m counters. Keep at most m keys with counts; on a key already present increment it; on a new key when fewer than m are held, insert at 1; otherwise decrement every held counter and drop those reaching zero. With m = 1,000,000 at roughly 40 bytes per counter, that is about 40 MB, comfortably inside 200 MB alongside the heap.
State the guarantee as a number, not as 'approximate'. Misra-Gries returns an estimate satisfying f - n/(m+1) <= f_hat <= f, so with n = 2e8 and m = 1e6 every count is an undercount by at most 200. If the 200th hottest pair has hundreds of thousands of writes, a 200-count error cannot reorder the top of the list — and if the counts near rank 200 are within 200 of each other, the ranking there was never meaningful for the decision being made. The sketch also merges: run one per shard, combine, prune back to m, and the bound holds with n the combined length.
If exactness is genuinely required, argue for two passes rather than more memory: the Misra-Gries pass yields a candidate superset of size m, and a second pass counts only those candidates exactly in O(m) memory. That is the standard way to get an exact top-K under a memory bound, and the cost is one extra read of the day's rows.

Follow-up

The decision downstream is which keys get a single writer. Does a rank error at position 190 change that decision, and what does the answer tell you about how much accuracy to buy?
Write volume on one key is bursty — a promotion concentrates it into twenty minutes. Does a daily top-K find it, and what window would?
How do you keep the sketch across a worker restart without replaying the day?

Fold late out-of-order status events into a monotonic leg state

Hard · semilattice fold · event time · idempotency

Leg statuses form the chain planned < tendered < accepted < picked_up < in_transit < delivered, with exception and cancelled outside it. You ingest 300 million observation_event rows a day carrying subject_id (leg), observed_status, occurred_at, received_at, clock_offset_ms and dedupe_key; five percent arrive over an hour late and a tail arrives a day late. Write the merge that produces each leg's current status, prove the result depends only on the set of events and not their order, and state the per-event cost.

Approach

Define the per-leg state as a tuple ordered totally: (status_rank, occurred_at, source_priority, event_id), and define merge(a, b) as the maximum under that order. Rank leads; occurred_at and the rest only break ties. The proof is then one line: maximum over a total order is idempotent, commutative and associative, so the state space with merge is a join-semilattice and the fold over any permutation of the event set yields the same value. That is also why no global sort is needed — the O(n log n) disappears and the fold is O(1) per event, O(distinct legs) state.
Rank before time, deliberately. occurred_at is device-reported and sometimes wrong; if it led the comparison, the leg's status would be handed to whichever device has the most badly skewed clock. Ranking first means a bad clock can only reorder events within one rank. Correct occurred_at by clock_offset_ms where it was measured, and where the offset exceeds a threshold keep the observation but bar it from winning a tie — the event is still evidence, it just stops being authority.
Make the lattice total, because exception and cancelled are not points on the chain and max over a partial order is undefined for incomparable pairs. Model cancelled as an absorbing top element, and carry exception as a boolean flag merged by OR alongside the rank rather than as a rank of its own — otherwise an exception at rank 4 either blocks delivered at rank 6 or is erased by it, and both are wrong. A product of semilattices is a semilattice, so the order-insensitivity argument still holds over the tuple.
Deduplicate on dedupe_key even though the merge is idempotent. Idempotence covers a byte-identical redelivery; a partner replaying a window with a fresh event_id is not identical, and the tie-break would let the replay displace the original for no reason. Dedupe on the stable key, then fold — both halves are needed, and conflating them is the usual defect.
Partition the stream by subject_id so all events for one leg land on one partition: broker ordering guarantees are per-partition only, and while a commutative merge makes interleaving harmless, a concurrent read-modify-write on the same leg row is still a lost update. Where that cannot be guaranteed, make the write a compare-and-set on a version column, or a conditional update that only raises the state. Because the state is mergeable, shard maps combine cleanly and replaying the whole day converges to the same answer — which is what makes reprocessing safe.
Say where genuine regressions live: a delivered shipment that is refused and returns is a new leg, not a rank decrease. Monotonicity is only sound if real reversals are modelled as new objects, and keeping every superseded observation is what lets you answer the dispute later, so the fold discards nothing — it marks losers superseded.

Follow-up

An event arrives a week late for a leg already closed and invoiced. Does it change the status, the on-time metric, both, or neither?
Two events for one leg have the same rank and the same occurred_at from different sources. What decides, and is that decision stable across a replay?
How would you test order-insensitivity rather than assert it?

Trace a suspect lot forward to every shipped order line

Medium · graph traversal · genealogy · indexing

lot_edge(parent_lot_id, child_lot_id) has 400 million rows with btree indexes on both (parent_lot_id, child_lot_id) and (child_lot_id, parent_lot_id). lot_shipment(lot_id, order_line_id) maps lots to shipped lines. Given one suspect received lot, return every downstream order_line_id. The reachable subgraph can reach 2 million lots and 12 levels; the full edge table does not fit in memory. Give the traversal, its complexity in terms of the reachable subgraph, and what you do if the data contains a cycle.

Approach

Genealogy is a directed acyclic graph, not a tree: a production run consumes several lots and one lot splits across many shipments, so a node is reached by several paths and a tree traversal revisits it combinatorially. Breadth-first from the suspect lot over the (parent_lot_id, ...) index, with a hash set of visited lot ids, is O(V' + E') where V' and E' count only the reachable subgraph — 2 million 8-byte ids is 16 MB of payload and a few tens of MB with open-addressing overhead, which is why the visited set is affordable and the edge table's 400 million rows never enter memory.
Batch each level into one indexed read. Issue WHERE parent_lot_id = ANY($1) with a few thousand parents per call instead of a query per node: at 12 levels and 2 million nodes, per-node round trips dominate the wall clock entirely, and the recall's deadline is a real-world response time, not a query-plan aesthetic. The batch size is the tuning knob between round-trip count and per-statement planning cost.
Keep the visited set even though the schema says DAG. Rework and repack loops — material returned into an earlier lot — do occur, and an unguarded traversal on a cyclic edge set does not terminate; a set membership test is O(1) and turns a hang into a correct answer plus a data-quality alert. Also cap the traversal by level and by node count, and fail loudly on the cap rather than returning a partial set silently.
Join to lot_shipment once at the end over every visited lot id, not over the leaves and not per node. A lot is routinely both shipped and consumed: half a lot goes out on an order line while the other half is repacked into a child lot, so it has a lot_shipment row and outgoing edges at the same time and is not a leaf. Restricting the join to nodes with no outgoing edges drops exactly those order lines, which is under-recall — the one failure mode this trace is not allowed to have. One bulk join against an index on lot_id over the whole visited set is a single pass; per-node lookups multiply the round-trip problem by the number of visited lots.
State the semantics honestly: reachability answers which shipped lines could contain material from the suspect lot, and without propagating edge quantities it is a superset of physical containment. For a recall that is the correct direction of error — over-recall, never under-recall — and saying so is the point, because a trace that returns quickly and incompletely will be believed. The upstream question, which received lots are in a suspect shipped unit, is the same traversal over the reverse index, which is why both indexes exist.

Worked solution (30 min)

Draw a six-lot fixture containing a merge (two parents into one child), a split (one parent into three children), and one interior lot that is partly shipped on an order line and partly consumed into a child lot, then write down the expected reachable set and the expected order lines by hand.
Implement level-batched BFS with an explicit frontier list, a visited hash set, and a per-level batch read.
Add a cycle to the fixture and confirm the traversal terminates and flags it rather than hanging.
Bulk-join lot_shipment once over every visited lot id — the whole visited set, not the leaves — and count the queries issued, confirming it is one per level plus one.
Re-run with the frontier order reversed to confirm the output set is unchanged.

Expected result: Every lot downstream of the suspect appears exactly once in the visited set; the merge node is visited once, not twice; the partly-shipped interior lot contributes its order line even though it has children; the cyclic fixture terminates with the same reachable set plus a flagged edge; the query count is levels + 1, not proportional to node count.

Verify it yourself

The reachable set is a superset of the set you computed by hand, and equal to it on the acyclic fixture.
Restricting the final join to lots with no outgoing edges drops the partly-shipped interior lot's order line — the failing test that pins the over-recall rule to code.
Running the same traversal over the reverse index from a leaf returns the suspect lot.
Removing the visited set makes the diamond fixture return duplicates — a failing test that documents why the set is there.

Follow-up

A recursive CTE would express this in ten lines. What is its cost here if the index on the traversal column is missing, and how would you notice?
You need the quantity of suspect material in each downstream line, not just reachability. What has to be on the edge, and what does that do to the traversal?
The traversal must run while a production run is writing new edges. What does a consistent answer mean here, and what do you cut it at?

Attribute multi-stop freight cost without fanning it out

Medium · many-to-many joins · window functions · integer money

shipment_leg(leg_id, shipment_id, sequence_no, linehaul_cost_cents, accessorial_cost_cents, actual_arrive_at) joins to fulfilment_order_line through leg_order_line(leg_id, order_line_id, allocated_weight_g). One multi-stop truckload leg carries 40 order lines. The current report SUMs linehaul_cost_cents after joining all three tables and returns roughly forty times the real freight spend. Write the query that returns cost per order for one month, explain the fan-out in terms of the join's row count, and make the per-order amounts sum exactly to the leg totals in integer cents.

Approach

Name the cardinality: joining shipment_leg to leg_order_line produces one row per (leg, line) pair, so the result has SUM over legs of lines-per-leg rows rather than one row per leg. SUM(linehaul_cost_cents) over that result counts each leg's cost once per line it carries, which is why 40 lines gives 40 times the spend. The bug is not the SUM, it is aggregating a parent measure across a child-expanded row set.
Compute a share before aggregating anything. In one pass: leg.linehaul_cost_cents * b.allocated_weight_g / SUM(b.allocated_weight_g) OVER (PARTITION BY b.leg_id), which keeps the denominator local to the leg and needs no self-join.
Do the division in integer cents and settle the remainder deliberately. Floor each share, then distribute the leftover cents one each by descending remainder, ordered by a stable tiebreak such as order_line_id. Floating-point shares leave a residue that makes the freight accrual miss by a cent per leg, which costs more to investigate than the freight.
Aggregate to the order only after the share exists: GROUP BY order_id over the per-line allocated cents. Restrict the month on shipment_leg.actual_arrive_at, not on the order date, so a leg is attributed in the period it actually ran.
Index for the shape of the query: leg_order_line needs (leg_id) for the window partition and (order_line_id) for the reverse lookup, and shipment_leg needs an index on actual_arrive_at to bound the month without a full scan.
Apply the same treatment to accessorial_cost_cents separately. Accessorials such as a detention charge often belong to one stop rather than to the whole leg, so weight-sharing them across every line on the load misattributes cost even though the total ties out.

Worked solution (30 min)

Build a fixture with one leg at 100,003 cents carrying three lines weighted 1000 g, 1000 g and 1 g.
Write the naive join and observe the total reported as three times the leg cost.
Rewrite using the SUM(...) OVER (PARTITION BY leg_id) share, floor to integer cents, and add a largest-remainder pass ordered by remainder then order_line_id.
Aggregate to order level and compare the grand total against SELECT SUM(linehaul_cost_cents + accessorial_cost_cents) FROM shipment_leg for the same window.

Expected result: Per-order allocated cents whose total equals the sum of leg costs for the month exactly, with no leg double-counted and the 2-cent remainder on the fixture distributed deterministically rather than lost or duplicated.

Verify it yourself

The reconciliation total matches to the cent, not to within rounding.
COUNT(DISTINCT leg_id) in the result equals the leg count for the month, proving no leg was expanded.
Re-run with the line order shuffled and confirm the same order receives the same remainder cent.
Add a fourth line to the leg and confirm the leg total is unchanged while the shares redistribute.

Follow-up

A line is cancelled after the leg ran. Does its share of the freight disappear, move to the remaining lines, or stay as a cost with no revenue?
Weight is the wrong basis for a light bulky pallet. What would you use instead, and does the reconciliation check still hold?
The carrier invoice arrives three weeks later and differs from the rated cost. Where does the correction land, and does it re-run this allocation?

Decide whether available-to-promise is derived or stored

Hard · denormalisation · read path · cache staleness

Available to promise at one node equals inventory_position.qty in state on_hand minus the sum of allocation.qty in status held or committed for that key. The read sits inside checkout: 4,000 reads per second at peak, p99 budget 20 ms. inventory_position holds 40 million rows; during a promotion a single hot item-node pair carries about 2,000 open allocations. Decide whether to compute this per read or to maintain available_qty on the position row. Give both paths, the failure mode of each, and what the browse read may do that the allocate read may not.

Approach

Cost the derived read in tuples, not in plans. One index probe into the position is trivial; the aggregate over allocation touches one tuple per open claim on that key. Even if only a tenth of peak lands on the hot pair, 400 reads per second times 2,000 rows is 800,000 index tuples per second for one item at one node, and it gets worse exactly as the key gets hotter. A partial index on status IN ('held','committed') does not help, because the row count is the cost, not the selectivity.
Maintaining available_qty on the position row is nearly free on the contention axis, which is the point most candidates miss. The allocator already takes a row lock or a compare-and-set on that same position row, so the extra column costs no lock that was not already held, and the read collapses to a single primary-key probe well inside 20 ms.
It is not free on the correctness axis. available_qty becomes a second derived value that can drift, so it joins qty under the same reconciliation job with its own re-sum against the allocation table, and a drift alert on that comparison is the operational price of the read speed.
Split the two reads rather than arguing about one. The browse read may serve from a replica or a short-TTL cache and must err low, showing less than is there. The allocate read must be on the primary and inside the same transaction as the conditional update, because replication lag peaks during the write burst that made the key hot. Conflating them is what makes a replica read look safe in a benchmark.
Pick the TTL against the drain rate, with a number. If a promotion burns 2,000 units in a minute, a 30-second cache is a thousand units stale, so the displayed figure should be a banded signal such as in stock or low stock rather than an exact count that the allocate step will then contradict.
Do not denormalise availability across nodes into a single network number. A yes-somewhere answer cannot be allocated against, so sourcing will disagree with the page, and the customer sees a promise retracted after checkout rather than an honest out-of-stock before it.

Follow-up

Roll this out on a live system. How do both paths run side by side, what do you compare, and what result gates the cut-over?
The reconciliation job finds available_qty drifted by 3 on one key. What do you do, and what must you not do?
Lot-tracked items multiply the position rows per item-node. Does your answer change, and where does the read now spend its time?

Explain your approach to balancing performance requirements with safety and compliance constraints.

Medium · system design and architecture

Approach

Fix the scope first: who calls this, how often, and what they do when it fails.
Choose a partition key and say what query it makes expensive.
Name the failure you are designing for, then the recovery path.

Follow-up

What breaks first when traffic grows ten times?
What would you drop to keep the system up under load?

How would you design a system to monitor and report on pipeline integrity data in real-time?

Medium · system design and architecture

Approach

Choose a partition key and say what query it makes expensive.
Name the failure you are designing for, then the recovery path.
Fix the scope first: who calls this, how often, and what they do when it fails.

Follow-up

What would you drop to keep the system up under load?
What breaks first when traffic grows ten times?

How do you ensure the reliability and security of software that interfaces with critical physical infrastructure?

Medium

Approach

State your assumptions explicitly before working the problem.
Work from the requirement backwards to the design.
Clarify what is being asked and what a complete answer contains.

Follow-up

What assumption would you test first?
How would you know your answer was wrong?

Can you describe a complex technical project you led and the specific technologies you chose to solve the problem?

Medium

Approach

Say what you would check first and why it is the highest-information step.
State your assumptions explicitly before working the problem.
Work from the requirement backwards to the design.

Follow-up

What assumption would you test first?
How would you know your answer was wrong?

Design the error taxonomy for an inventory reservation API

Easy · error taxonomy · retries · api contract

The caller is a checkout service working inside a 50 ms budget; on any non-2xx it applies a generic three-attempt retry policy. Design the error contract for POST /reservations on the Availability and Inventory Service. Cover input validation, credentials, insufficient available quantity, a lost compare-and-set on inventory_position.version, quota exhaustion, and a write whose outcome is unknown after a timeout. For each class give the status code, a stable machine code, whether a retry is safe, and whether that retry reuses the idempotency key. Deliver the envelope schema and the classification table.

Approach

Fix the envelope before the list: application/problem+json (RFC 9457, which obsoletes RFC 7807) with type, title, status, detail and instance, plus two extension members - a stable code and a retry class of never, after_backoff or same_key. Callers branch on the code; the prose stays free to change.
Separate a business answer from a transport failure. Insufficient availability is a correct, computed answer, so it belongs in the reservation representation as 200 with requested_qty, available_qty and a reason, not behind a 5xx. Dressing it as a server error points every retry layer in the caller's stack at a hot item-node row that will not have stock for hours.
Classify the rest on two axes - is the resulting position known, and is the caller at fault. Known and caller's fault: 422 for an unknown SKU or a UoM conversion that does not divide exactly, 401/403 for credentials, never retryable. Known and ours: 409 for a lost version CAS, retryable immediately with the same key under a bounded attempt count, because that is contention rather than a bad request. Unknown: 502/504, retryable only with the original key. Overload: 429 with Retry-After.
Write the unknown rule explicitly, because it is the only class where the caller can do damage: the movement may have committed, so the two permitted actions are replay with the identical Idempotency-Key or resolve with GET /reservations?idempotency_key=..., and issuing a fresh reservation is forbidden.
Pin the taxonomy with a test asserting every error path emits a code from the enumerated set. Without it the catch-all handler ships an uncoded 500, and the caller's switch falls through to a default branch that is almost always retry.

Worked solution (20 min)

Write the problem+json schema with its two extension members: code (stable enum) and retry_class (never | after_backoff | same_key).
Enumerate ten concrete failures from the real path: unknown item_id, closed node, half-case quantity, expired credential, key reused with a different body, duplicate key still in flight, quota exceeded, version CAS lost, position store timeout, available_qty below requested_qty.
Place each in the table with status, code, retry_class and the caller's one-line action.
Re-read the availability rows and confirm they are 200 responses inside the reservation representation rather than entries in the error envelope.
Write the caller's switch in pseudocode and check that every branch is reachable from the table and nothing falls through to a retrying default.

Expected result: A table in which insufficient availability is a 200 carrying available_qty, exactly three classes are retryable - 429, the version conflict, and the indeterminate 5xx - every retryable row that mutates state states the original Idempotency-Key is resent unchanged, and no other 4xx is marked retryable.

Verify it yourself

Every error carries a code distinct from its HTTP status, so two different 409s are separable by a caller that never reads the prose.
A caller implementing only the retry_class column cannot retry a validation failure or a stock shortfall.
Replaying a completed idempotency key returns the same status code as the first call, including when that code was a 4xx.

Follow-up

A 409 from a lost version CAS and a 409 from a key replayed with a different body demand opposite caller behaviour. What in the contract tells them apart?
Your p99 under contention is 120 ms against a 50 ms caller budget. What do you return at the deadline, and which retry class does it carry?
How do you add a new error code later without breaking a client that switches exhaustively over the current set?

Duplicate bookings and second trailers after a partner slowdown

Hard · idempotency · partial failure · retries

A partner's booking API ran at 45 s p99 for ninety minutes. Afterwards 214 shipment_leg rows carry two booking references, 41 stops received two trailers, and finance holds duplicate invoices. booking_idempotency_key is generated and persisted on the leg before the call and is sent to the partner, whose documented dedupe window is 60 seconds. The HTTP client retries timeouts three times, and a sweeper re-tenders any leg still in status 'planned' after ten minutes. Explain how duplicates happened despite the key, and give the corrected flow.

Approach

Establish that the duplicates are timeouts rather than double-clicks. Join the gateway request log to shipment_leg for the 214 legs and check whether each duplicate pair carries the same idempotency key and how far apart the calls landed. Same key, calls minutes apart, first attempt ending in a client timeout is the ambiguous-outcome signature; different keys would mean the key is being regenerated and is a different bug.
Name why the key did not help. The key was correct and persisted, but deduplication was delegated to the partner, whose window is 60 seconds. A retry at 45 s may land inside it; the third retry does not, and the ten-minute sweeper re-tender certainly does not, so the partner sees a new request and books a second truck. A short partner window plus a long retry horizon is an at-least-once system wearing an at-most-once label.
Identify the missing state. The leg goes planned to tendered only on a 2xx, so a timeout leaves it exactly as if nothing was sent, which is false: the partner may have committed before the connection dropped. The state machine needs an explicit in-doubt status between requested and confirmed, and nothing may issue a new tender for a leg in that state.
Replace re-issuing with asking. On any ambiguous outcome, move the leg to in-doubt and let a sweeper resolve it by querying the partner for that key or for bookings against our shipment reference, and only tender again when the query definitively returns nothing. Where the partner offers no query, the manifest or the status feed is the authority and resolution is slower, so the in-doubt state must be allowed to persist rather than being timed out into a retry.
Move retry ownership out of the process. A retry decorator around an HTTP client disappears on deploy or crash, so the persisted state is the only thing that can drive recovery: a sweeper reading legs in requested or in-doubt, with bounded attempts, backoff and an alert on in-doubt depth. The call site should not retry an irreversible effect at all.
Deal with the 214 that already exist, since code cannot undo them. Reconcile against the partner's manifest, cancel the duplicate where the cancellation window still allows it, and for legs already executed record the second cost as a compensating entry and dispute the invoice. State plainly that this is a business process with a cost, which is exactly why the guard belongs before the call.

Follow-up

The partner offers no lookup by your key and no search by reference. What is your resolution path and how long may a leg stay in doubt?
Which other effects in this system have the same shape and the same missing state?
How do you test this without a degraded partner, and what does the fault injection have to simulate that a plain timeout does not?

Suggested practice checklist

Day one measures instead of guessing, under a fixed rubric, and the remaining hours are allocated in proportion to the gaps before any studying begins. The allocation is deliberately not renegotiated midweek, because the area that feels worst on day three is usually the one that is moving.

Day 1 — Diagnostic, scored before you study anything

Sit a 110-minute diagnostic in four blocks: forty-five minutes on two coding problems, twenty-five on one design prompt taken to interface and data model, twenty of short-answer fundamentals, and twenty delivering two behavioural answers aloud.
Score each block from 0 to 3 on a fixed rubric where 3 is correct and fluent, 2 is correct but slow or prompted, 1 is partially correct and 0 is stuck, grading the artifact rather than how the attempt felt.
Allocate days two to five in proportion to 3 minus each block's score, write the allocation down, and commit to leaving it alone.

Deliverable: A scored rubric and a fixed hour allocation for the rest of the week.

Day 2 — Largest gap: find the boundary rather than the subject

Split the weakest area into named sub-skills and rate each separately. For coding those are restating the problem, choosing the structure, stating the invariant, turning the invariant into loop bounds, handling empty and single-element input, and accounting for complexity out loud.
Attempt three items positioned just above where the rating drops off, and for each write the first move you failed to make.
Re-attempt one of them from blank four hours later with nothing open.

Deliverable: A sub-skill map with the two blocking sub-skills circled.

Day 3 — Drill the blocking sub-skill by repeating the shape

Do eight short repetitions of the same shape rather than eight different problems, so what gets practised is the pattern and not the puzzle.
State the rule you now hold in one sentence, then test it against a case built to break it, a sliding window over an array containing negative values, or a cache-aside read path whose invalidation message is dropped.
Have someone else read your one-sentence rule and find the precondition you left out.

Deliverable: One rule statement with its preconditions attached and one counterexample that would have caught the incomplete version.

Day 4 — Second gap, plus maintenance on the strongest area

Run the same sub-skill decomposition on the second-largest gap in half the time.
Spend twenty-five timed minutes on the block you scored highest, choosing the hardest item you can still finish rather than a warm-up.
Write whether each area fails you on recall, on setup, or on execution, and set the fix accordingly: repetition for recall, a written checklist for setup, timed work for execution.

Deliverable: A second sub-skill map plus a one-line failure diagnosis for each area.

Day 5 — The gap that is not a skill

Record one technical and one behavioural answer, then count two things in the playback: seconds before your first clarifying question, and sentences you began without knowing where they would end.
Practise saying that you do not know, followed by how you would find out, without letting it soften into a guess, and practise stating a complexity or an estimate before being asked for it.
Redeliver one answer under a hard ninety-second cap, which forces structure ahead of detail.

Deliverable: Two recordings with a counted reduction in time-to-first-question.

Day 6 — Retest under day-one conditions

Sit the same 110-minute structure with new prompts of comparable difficulty and score it on the identical rubric.
For any block that did not move, change the method rather than adding hours: a block stuck at 1 usually means the practice was too varied, not too short.
Write down which single block you would still lose the offer on.

Deliverable: A second scored rubric placed beside the first, with one named remaining risk.

Day 7 — Full loop under interview conditions

Run a sixty-minute mock over the two blocks that moved least, with an interviewer briefed to interrupt and change direction mid-answer.
Write the recovery script for going blank: restate the question, state your assumption, name the first thing you would check.
Say every rule from the week aloud without reading it, and cut any you cannot state in a single sentence, since a rule you have to reconstruct mid-answer will not survive an interruption.

Deliverable: A one-page card holding the recovery script and only the rules you could state from memory.

The estimate and the day you knew

A slipped date is only a bad story if you sat on it. What matters is what you believed when you gave the number, the signal that told you it was wrong, how many days passed before you said so, and what you cut rather than asking for more time. Scope you defended counts as much as scope you dropped.

What is your experience with real-time data processing or latency-sensitive applications?

Medium · behavioural and engineering judgement

Approach

State the situation in two sentences and spend the rest on the reasoning.
Name the disagreement and how you resolved it with evidence.
Close with what you would do differently, concretely.

Follow-up

What did you decide not to do, and why?
How did you know your change caused the improvement?

Describe a situation where you identified an inefficiency in a process and what steps you took to improve it.

Medium · behavioural and engineering judgement

Approach

Pick a story where you made the decision, not one where you watched it.
Name the disagreement and how you resolved it with evidence.
Close with what you would do differently, concretely.

Follow-up

What would you do differently if you ran that again?
How did you know your change caused the improvement?

Tell me about a time you had to explain a complex technical concept to a non-technical stakeholder.

Medium · behavioural and engineering judgement

Approach

Pick a story where you made the decision, not one where you watched it.
State the situation in two sentences and spend the rest on the reasoning.
Close with what you would do differently, concretely.

Follow-up

What did you decide not to do, and why?
What would you do differently if you ran that again?