Interview guides · Perceptive

Perceptive Machine Learning Engineer Interview Questions & Guide 2026

Perceptive Machine Learning Engineer interview guide: reported rounds, practice questions with worked solutions and a 7-day prep plan.

The 60-second brief

As a Machine Learning Engineer at Perceptive, you are at the forefront of transforming medical imaging through advanced artificial intelligence. This role is not merely about building models; it is about architecting high-stakes, mission-critical solutions that directly impact diagnostic accuracy and patient outcomes. You will be tasked with solving complex problems in a domain where precision is non-negotiable and innovation drives the standard of care.

Seniority moves the scope further than the words in the title do. An earlier-career loop mostly checks that you implement something correctly and can reason about its cost, while a senior loop checks that you can pick between two defensible designs and say what you gave up.

Focus areas

Bound every outbound call with a timeout
Make every write idempotent under retry
Detect concurrent edits instead of losing writes

Overview

You will operate at the intersection of Computer Vision, Medical Imaging, and Software Engineering. The position requires a unique blend of deep academic understanding and the ability to deploy scalable, robust production code. Whether you are working on Senior AI / ML Engineer initiatives or taking on the strategic responsibilities of an Associate Director, AI/ML Engineering, your work will define the next generation of Perceptive technology.

Perceptive prioritizes candidates who can demonstrate a balance between rigorous scientific research and pragmatic, scalable software engineering practices.

Where to focus your preparation

Treat the sections below as preparation areas and confirm the format with your recruiter.

Preparation focus

No round sequence has been reported for this company, so work the categories below and confirm the format with your recruiter.

What to demonstrate

Breadth across SQL, experimentation and product reasoning
Ability to state assumptions before choosing a method

How to prepare

Drill the practice exercises below and time yourself
Prepare three quantified stories about decisions you drove

Where candidates lose points

Running a schema change as though the lock lasts as long as the statement

In PostgreSQL an ALTER TABLE that needs an ACCESS EXCLUSIVE lock must first wait for every open transaction touching that table, and while it waits, later queries needing a conflicting lock queue behind it rather than overtaking it. A DDL statement that would execute in milliseconds, issued while a thirty-second analytics query is open, therefore stalls all traffic on that table for thirty seconds: the outage length is set by the longest open transaction, not by the change. The defences are specific and worth knowing by name - set lock_timeout low and retry rather than queue, add columns without a volatile default so no table rewrite occurs (from version 11 a non-volatile default is a metadata-only change), build indexes with CREATE INDEX CONCURRENTLY while accepting that it cannot run inside a transaction block and leaves an invalid index behind if it fails, and add constraints as NOT VALID followed by a separate VALIDATE CONSTRAINT, which takes a weaker lock.

Letting a slow dependency consume unbounded concurrency

The failure that takes a service down is usually not an error but a delay. A dependency answering in thirty seconds instead of fifty milliseconds holds each request's worker or connection six hundred times longer, and since required concurrency is arrival rate times latency, a fleet sized for sixty in-flight requests now needs thirty-six thousand to sustain the same rate - so it queues, and requests whose clients have already abandoned them still occupy resources. Retries make it precisely worse: a policy of three attempts triples the load on a dependency at the exact moment it is least able to serve, which is how one slow dependency becomes an outage of everything sharing that pool. Containment is four specific things - a timeout on every outbound call shorter than the caller's remaining budget, a bounded pool per dependency so one cannot starve the others, backoff with full jitter rather than a fixed delay so retries do not resynchronise, and a circuit that stops sending once the failure rate makes an attempt pointless.

Check-then-act on shared state

Read, decide, write is not safe under concurrency unless the decision and the write are one atomic step: a unique constraint with conflict handling, a compare-and-set, or a row lock held for the whole transaction. Two requests can both pass the existence check before either inserts, which shows up as duplicate rows under load and never in a single-threaded test.

Trusting input because it came from your own front end

Anything crossing a trust boundary is hostile: parameterise queries instead of building SQL by concatenation, validate against an allow-list rather than a deny-list, and bound the size of anything you allocate from a request. Raising this unprompted in an API or design question is a cheap and unusually strong signal.

Questions & practice

Describe your process for validating a model’s performance in a high-stakes medical context.

Medium · machine learning fundamentals

Approach

Say how you would validate it, and where leakage could enter the split.
Name the simplest model that could work and what would make you move past it.
Pick the metric from the cost of each error type, not from habit.

Follow-up

How would you know the model is overfitting?
What changes if the classes are heavily imbalanced?

How do you ensure your model generalizes well across different imaging hardware or clinical sites?

Medium · machine learning fundamentals

Approach

State the learning problem: the label, the unit of prediction and how the model is used.
Say how you would validate it, and where leakage could enter the split.
Pick the metric from the cost of each error type, not from habit.

Follow-up

What changes if the classes are heavily imbalanced?
How would you know the model is overfitting?

Track a rolling failure rate per destination for circuit decisions

Easy · sliding window · ring buffer · circuit breaker

The egress service delivers about 1,500 webhooks per second across roughly 40,000 destinations, each call bounded by a 10 second timeout. Maintain, per destination, the failure rate over the trailing 60 seconds so a caller can ask before dispatch whether the circuit should open. Attempts arrive as (destination_id, finished_at_ms, outcome). Requirement: amortised O(1) per attempt, with total memory bounded by the destination count rather than by traffic. Give the structure, its exact memory, and the rule that stops a destination with three attempts from opening a circuit.

Approach

Name the exact-deque version and then reject it as the default. Holding timestamps and advancing a tail pointer past anything older than now minus 60 seconds is a correct two-pointer window at amortised O(1) per attempt, but its memory tracks in-window traffic, so one destination in a retry storm holds hundreds of thousands of entries while thousands of quiet destinations hold none.
Use a ring of 60 one-second buckets per destination, each bucket a pair of counters for attempts and failures. On an attempt, advance the ring by the elapsed whole seconds, zeroing at most min(elapsed, 60) buckets, then increment the head. That is amortised O(1) with a fixed footprint per destination.
State the footprint: 60 buckets times two 4-byte counters is 480 bytes of payload per destination, so 40,000 destinations is roughly 20 to 25 MB with per-entry overhead, bounded by the catalogue rather than by the rate. The cost is granularity, since the oldest bucket ages out in whole seconds, which is far tighter than the decision needs.
Require a minimum sample before the circuit may open. A destination with three attempts and three failures reads as 100 percent and is not evidence; a floor of roughly 20 attempts in the window makes the ratio meaningful, and below that floor use a run of consecutive failures as the trigger instead.
Expire idle destinations, or memory grows with every destination ever seen rather than with the live set. Hold the rings in a bounded LRU keyed on destination_id and treat a miss as no history, which is the correct default for an endpoint that has been silent for a minute.
Keep the half-open probe out of the window arithmetic. After the circuit opens, one probe per interval decides whether to close it, and folding that single success into a window that still holds a 100 percent failure history would reopen the destination on one data point.

Worked solution (20 min)

Define the bucket struct and the advance step: take floor(finished_at_ms / 1000), compare with the ring's current second, zero min(delta, 60) buckets forward, then write into the new head.
Trace a destination that receives 5 attempts, goes silent for 90 seconds, then receives one more, and confirm the rate is computed from one attempt rather than six.
Compute total memory for 40,000 destinations at 60 buckets of two 4-byte counters, and state what changes if the window widens to 300 seconds.
Write the open rule as a single predicate combining the minimum-attempt floor with the rate threshold.

Expected result: A per-destination ring of 60 one-second (attempts, failures) buckets advanced lazily for amortised O(1) cost, roughly 480 bytes of counters per destination and about 20 to 25 MB for 40,000 of them, an LRU bound on live rings, and an open rule requiring both a minimum attempt count in the window and a rate above threshold.

Verify it yourself

After 60 seconds of silence a destination's window holds zero attempts and cannot open a circuit.
Memory measured with one hot destination and 39,999 idle ones is indistinguishable from the reverse, which the deque version would fail.
Replaying the same attempt stream with the clock shifted by an hour yields the same final rate.

Follow-up

The fleet is 30 instances and each sees roughly a thirtieth of a destination's traffic. Where does the rate actually live, and what does a per-instance answer get wrong?
A destination answers in 9.5 seconds and succeeds. It is not failing but it is consuming your per-destination concurrency. What signal should open the circuit here?
How would you make the window survive a process restart, and is it worth the cost?

Denormalise tenant onto revisions and backfill it live

Hard · denormalisation · online migration · backfill

resource_revision (revision_id, resource_id, version, actor_user_id, change_kind, patch, request_id, created_at) has 400M rows and no tenant column; tenant_id lives only on resource. Two reads need it: a tenant-scoped audit feed ordered by created_at DESC, and an offboarding purge. Both join back to resource today. Justify adding tenant_id to resource_revision against those two reads, name the anomaly the copy introduces and the constraint that prevents it, then give the ordered migration for a live table taking 1.2k writes/second — the lock each step takes, how the backfill is batched, and where each step stops being reversible. PostgreSQL 16.

Approach

Justify from the access path rather than from taste. Without the column, the audit feed either scans resource_revision by created_at and discards other tenants' rows, or resolves the tenant's resource_ids first and probes with them — both proportional to the tenant's whole history rather than to one page. With (tenant_id, created_at DESC, revision_id DESC) it is a seek that stops at 50 rows, and the purge becomes a ranged delete instead of a join.
Name the cost exactly: a second copy of a fact can disagree with the first. Make the disagreement unwritable rather than documented — add UNIQUE (resource_id, tenant_id) on resource so it can serve as a foreign-key target, then FOREIGN KEY (resource_id, tenant_id) REFERENCES resource (resource_id, tenant_id) on the revision table. A revision can then only ever carry its parent's tenant.
Step one, expand: ALTER TABLE resource_revision ADD COLUMN tenant_id BIGINT NULL, with no default, so it is a catalogue change and no rewrite. It still needs ACCESS EXCLUSIVE for an instant, and that instant queues behind the longest open transaction on the table while every later query queues behind it — set lock_timeout to 2s and retry rather than wait.
Step two, dual-write: deploy the writer that populates tenant_id on every new revision while reads still use the join. Reversible by redeploying the previous build, because nothing reads the column yet.
Step three, backfill: batch by primary key rather than by created_at so the cursor is dense and resumable — UPDATE resource_revision rr SET tenant_id = r.tenant_id FROM resource r WHERE r.resource_id = rr.resource_id AND rr.revision_id > $1 AND rr.revision_id <= $1 + 5000 AND rr.tenant_id IS NULL — committing per batch and persisting the cursor. Throttle on replica replay lag and on dead-tuple count, since each batch writes 5,000 new row versions. Run the backfill before the index exists so those updates can stay HOT.
Step four, index then enforce then contract: CREATE INDEX CONCURRENTLY (cannot run inside a transaction block, scans the table twice, waits on open transactions, and leaves an INVALID index to drop concurrently if it fails); ADD CONSTRAINT ... CHECK (tenant_id IS NOT NULL) NOT VALID, then VALIDATE CONSTRAINT, which takes only SHARE UPDATE EXCLUSIVE, after which SET NOT NULL uses the validated check instead of re-scanning on PostgreSQL 12 and later. Only then move the audit reads onto the column and, in a later deploy, delete the join path.

Follow-up

The backfill is half finished and a rollback is required. What state is the table in, and what does the previous build do with a half-populated column?
How do you verify the backfill actually finished, given rows are still being inserted while it runs?
A resource must now be movable between tenants. What does that do to the composite foreign key and to the revisions already written?

Find version gaps and relay lag with window functions

Medium · window functions · event tables · frame semantics

outbox_event holds event_id, aggregate_type, aggregate_id, aggregate_version, event_type, payload, status ('pending','published','dead'), attempts, created_at, published_at. A projection is missing rows and you must decide whether the relay skipped events or the consumer dropped them. Write three queries over the last seven days: one listing every aggregate_id whose published aggregate_version sequence has a hole, one giving per-day counts with a running total, and one returning the newest published event per aggregate. For each, say where the window function is evaluated relative to WHERE and LIMIT. PostgreSQL 16.

Approach

Gaps: compute lead(aggregate_version) OVER (PARTITION BY aggregate_id ORDER BY aggregate_version) in a subquery, then filter next_version <> aggregate_version + 1 in the outer query. Window functions are evaluated after WHERE, GROUP BY and HAVING and before the outer ORDER BY and LIMIT, so the predicate cannot sit in the same WHERE clause and PostgreSQL 16 has no QUALIFY.
Say what the seven-day filter does to the answer: it truncates every partition, so the first row per aggregate has no predecessor inside the window and a hole spanning the boundary is invisible. Widen the window, or join to resource.version as the authority for the true maximum.
Running total: SELECT date_trunc('day', created_at) AS d, count() AS n, sum(count()) OVER (ORDER BY date_trunc('day', created_at) ROWS UNBOUNDED PRECEDING). An aggregate inside a window call is legal because grouping runs before windowing. The grouping key is unique per row here so ROWS and RANGE agree, but write the frame anyway — over ungrouped rows with tied timestamps the default RANGE frame pulls in every peer row and the total jumps.
Newest per aggregate: DISTINCT ON (aggregate_id) ... ORDER BY aggregate_id, aggregate_version DESC is the cheap PostgreSQL-only form when an index matches that order; row_number() OVER (PARTITION BY aggregate_id ORDER BY aggregate_version DESC) = 1 is the portable form and needs a subquery for the same evaluation-order reason as the gap query.
Interpret rather than report: no gaps plus a normal p95 of published_at - created_at points at the consumer; gaps or a fat lag tail point at the relay; rows still 'pending' with attempts > 0 point at neither, because they never left the database.
Be explicit that the partial index on (created_at, event_id) WHERE status = 'pending' does not serve any of these — they read published rows. Name the index a recurring monitor would need, and say why a query run twice a year may not deserve one.

Worked solution (30 min)

Write the three queries against seven days of data and confirm each returns without error.
In a scratch copy, delete one middle event for a single aggregate and confirm the gap query names that aggregate and the versions either side.
Run a running total over ungrouped rows ordered by date_trunc('second', created_at), once with the default frame and once with ROWS, and record where the two series diverge.
Compare the DISTINCT ON and row_number() plans on the same data and record rows-read for each.

Expected result: The gap query returns one row per hole with the version on each side; the daily query returns seven rows whose running total ends at the seven-day count; the newest-per-aggregate query returns exactly one row per aggregate_id present in the window. The RANGE running total is higher than the ROWS one at every tied second.

Verify it yourself

The per-day counts sum to COUNT(*) over the same predicate.
Re-running the gap query after reinserting the deleted event returns zero rows.
count(*) of the newest-per-aggregate result equals count(DISTINCT aggregate_id) over the same filter.

Follow-up

Relay failover redelivers events. Does a duplicate break the gap query, and how would you detect one from this table alone?
Turn the gap check into a continuous monitor rather than a query someone runs after an incident. What does it watch?
The consumer claims it never received event 4,812,006. What do you look at, in what order?

How would you design a pipeline for real-time inference on large medical images?

Medium · ml system design and production

Approach

Say where features come from at serving time and how they match training.
Separate the offline training path from the online serving path.
Fix the product goal and the online metric before choosing any model.

Follow-up

What happens when a feature is missing at serving time?
How would you detect drift before the metric drops?

Explain the trade-offs between different architectures for image segmentation.

Medium · system design

Approach

Choose a partition key and say what query it makes expensive.
Name the failure you are designing for, then the recovery path.
Name the read and write paths separately; they rarely have the same bottleneck.

Follow-up

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

How do you approach data augmentation when dealing with volumetric medical data?

Medium

Approach

Say what you would check first and why it is the highest-information step.
State your assumptions explicitly before working the problem.
Clarify what is being asked and what a complete answer contains.

Follow-up

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

Keep one unresponsive destination from stalling all webhook delivery

Medium · fan-out · circuit breaker · timeouts · bulkheads

Egress delivery sends about 1.5k webhooks/second to 40k destinations, with a per-destination concurrency cap of 4 and a 10-second connect-plus-read timeout. One destination begins accepting connections and never responding; within the hour 150 destinations behave the same way. Design the delivery path so unrelated destinations are unaffected: the pool structure, the timeouts, the retry policy, the per-destination circuit, and what is recorded so a retry is not a second effect at the receiver. State how many in-flight slots the degraded destinations hold and why that number decides the design.

Approach

Start with the number, and with the law that produces it. In-flight work is arrival rate times time in service, so 1.5k/second against a healthy 200 ms response needs about 300 concurrent slots. Per destination the same product applies, ceilinged by the concurrency cap: at the fleet average of 0.0375 deliveries/second per destination (1.5k spread over 40k) a 10-second timeout is 0.375 slots. A destination that has queued retries behind it is a different regime - every slot refills the instant an attempt expires, so it sits pinned at its cap of 4 - and 150 of those hold 600 slots, more than a pool sized for healthy traffic, entirely consumed by endpoints that will never answer. The per-destination cap bounds one endpoint and says nothing about the aggregate, which is exactly why it alone is not containment.
Contain with bulkheads and an admission bound rather than a larger pool. Cap total in-flight per pool and shard destinations across pools by a hash of destination id, so a correlated group - one provider, one region - cannot exceed its pool's share. A delivery refused admission and re-queued with backoff is strictly better than one holding a slot on behalf of a receiver that is not listening.
Treat the timeout as two timeouts, and be exact about what shortening one buys. Connect and read are separate failures and both must be shorter than the budget of whatever is waiting. Occupancy is min(cap, arrival rate x timeout), so dropping the read ceiling from 10 seconds to 3 cuts a merely slow destination's occupancy proportionally, 0.375 slots to 0.11 at the fleet average. It does not cut the 4 slots held by one of the 150: a destination with a retry backlog arrives far above cap/timeout - 0.4/second at a 10-second timeout, 1.33/second at 3 - so it stays pinned at the cap either way and only the slot-seconds per attempt fall. What that does buy is detection rate: 3.3x more failures observed per second on the same four slots, which is how fast the circuit reaches its threshold. Pick the value from the measured latency distribution of successful deliveries, with their high percentile as the floor, not from a round number.
Add a circuit per destination, counting a timeout as a failure. Once open, fail fast without taking a slot - that is the whole point, converting 4 held slots into zero. Half-open on a schedule with exactly one probe and close only if the probe succeeds, so a permanently dead endpoint costs one request per interval instead of a growing retry queue.
Make retries safe and non-synchronising. Back off with full jitter, sleeping a random value in [0, min(cap, base x 2^attempt)], because a fixed delay re-synchronises every failed delivery to one destination into a simultaneous burst. Delivery is at-least-once, so the payload carries the event id under the signature and the receiver deduplicates on it; record the attempt against (destination, event id) rather than a bare success flag, so a lost response does not become a second business effect on the other side.

Worked solution (25 min)

Compute healthy in-flight from rate times latency, then slots held by 150 destinations at the cap and the full timeout, and compare both against one pool size.
Write the pool sharding rule and the admission bound, and state what a refused delivery does next.
Pick connect and read timeouts from the success-latency distribution, then compute min(cap, arrival rate x timeout) for an average destination and for one with a retry backlog, and say which of the two the shorter timeout actually moves.
Write the circuit's state machine with its open threshold, probe interval and close condition, and the backoff formula with full jitter.

Expected result: Destinations sharded across bounded pools with a total in-flight cap per pool, separate connect and read timeouts chosen from the success-latency distribution, a per-destination circuit that fails fast while open and probes once per interval, and full-jitter backoff on retry. A shorter read timeout lowers occupancy only for destinations arriving below cap/timeout; a dead destination with a retry backlog stays pinned at its cap of 4 whatever the timeout, and the circuit opening is what takes it to zero. Deliveries carry a signed event id the receiver deduplicates on, and attempts are recorded per (destination, event id). Degraded destinations hold at most their pool's share rather than 600 slots of a shared pool.

Verify it yourself

Slots held by degraded destinations are bounded by the pool shard, not by how many destinations degrade.
An open circuit reduces slots held by that destination to zero between probes.
Lowering the read timeout does not reduce successful deliveries, verified against the measured success-latency distribution, and does not by itself reduce the slots a backlogged dead destination holds.
A redelivered event is discarded by the receiver on its event id rather than applied twice.

Follow-up

The destination is not dead - it answers in 9.5 seconds with a 200. Does a failure-rate circuit open? Should anything shed that traffic, and on what signal?
One destination requires deliveries in order. What does a per-destination concurrency of 4 do to that guarantee, and what would you change to offer it?
A destination has been parked six hours with 900k undelivered events. What does resuming look like, and is delivering the whole backlog the right call?

One customer endpoint stalls deliveries to every other destination

Medium · bulkhead · timeouts · saturation · retry amplification

The egress service delivers about 1.5k webhooks/second across 40,000 destinations, with a per-destination concurrency cap of 4 and a 10-second connect-plus-read timeout. Throughput falls to 300/second, queue depth climbs, and p99 delivery latency for unaffected destinations goes from 200 ms to minutes, while the error rate barely moves. One tenant holds 900 destination rows whose URLs share a hostname that now answers in 9.5 seconds. Explain the mechanism with the arithmetic, then give the containment in the order you would apply it.

Approach

Look at saturation before errors. A flat error rate with collapsing throughput says nothing is failing, things are waiting, so the first signal to pull is in-flight request count or pool wait time rather than the error counter. This is the distinction that decides the whole investigation.
Group in-flight work by resolved host, not by destination id. The cap is keyed per destination row, so 900 rows sharing one hostname buy 3,600 concurrent slots against a single host, each held for 9.5 seconds. The bulkhead was never a bulkhead for that host, and grouping by the wrong dimension is why the dashboard looked healthy.
Do the arithmetic in both directions. Required concurrency is arrival rate times latency, so 1.5k/second at 200 ms needs about 300 in flight, which is entirely consumed by 3,600 slow slots; conversely whatever concurrency is left sustains rate equals concurrency divided by 9.5 seconds, which is the 300/second you are seeing. Matching both numbers is what promotes this from a plausible story to the mechanism.
Explain why the circuit breaker never helped. It opens on consecutive failures, and a 9.5-second response inside a 10-second timeout is a success. Slow is not failing, so an error-rate breaker cannot see this; you need a slow-call ratio, a deadline propagated from the caller's remaining budget, or a concurrency limiter.
Contain in order: park the offending host so the shared pool drains, add a per-resolved-host concurrency cap alongside the per-destination one, give slow hosts their own queue so they cannot occupy the general pool, derive the timeout from the delivery deadline rather than a round number, and check the retry policy is not tripling load on a host that is already slow. Use backoff with full jitter so retries do not resynchronise on recovery.
State the invariant you are restoring: one tenant's endpoints degrade only that tenant's deliveries. That is a property to load-test for, not to assume from a config value.

Follow-up

The host recovers to 80 ms. How long does the queue take to drain, and what does the drain do to the recovered host?
Where should the 10-second timeout number actually come from?
If that tenant had one destination row instead of 900, would the cap of 4 have saved you? What would you measure to be sure?

Suggested practice checklist

For someone fluent in a dynamic language who has shipped real work but has never had to say what the runtime is doing underneath. The week is built on measuring and deliberately breaking things, because the questions that expose this background are the ones where the interviewer asks why a second time.

Day 1 — Measure before reasoning

Take a slow piece of your own code, write down in advance where you believe the time goes, then profile it and record how wrong the guess was. The cost is usually an allocation you did not notice or an accidental quadratic membership test.
Replace one list membership test inside a loop with a set and measure at a thousand, ten thousand and a hundred thousand elements, confirming the shape of the curve rather than only that it got faster.
Write down the three quantities you can now measure instead of assert: wall time, peak memory, and call count for the function you suspected.

Deliverable: A before-and-after profile of real code plus a written note on the size of the gap between the guess and the measurement.

Day 2 — References, copies, and the bugs they produce

Write the function with a mutable default argument, call it three times, and explain the accumulating result: the default is evaluated once when the function is defined, so every call shares one object.
Build a nested structure, take a shallow copy, mutate an inner element, and show that both views changed, because a shallow copy duplicates the container and not the elements. Then fix it with a deep copy and state the cost you just accepted.
Write two functions, one mutating its argument in place and one rebinding the local name, and predict the caller's view of each before running it. That single distinction produces most of the bugs that pass their tests.

Deliverable: Three small programs whose output you predicted correctly before running, each with a one-line statement of the rule underneath.

Day 3 — Types, once, in a language that checks them

Port one module you have already written, roughly a hundred lines, into a statically typed language, and record every place the compiler demanded an answer your original had left implicit: a value that can be absent, a numeric width, a case never handled.
Write the same signature in both languages and state what the static one guarantees before the program runs and what it does not, since it will not save you from a wrong algorithm or an index out of range.
Write the difference between an interface satisfied by declaration and one satisfied structurally, with one case each where the other approach would miss the mistake.

Deliverable: One module in two languages plus a list of the questions the type checker forced you to answer.

Day 4 — Concurrency, starting with what actually runs at the same time

Run the same CPU-bound function across four threads and four processes and measure both. Under the default CPython build the threaded version will not speed up, because only one thread executes bytecode at a time; the process version will. Check which build you are on first, since free-threaded builds remove that lock and change the result.
Then run a blocking I/O workload across four threads and measure it speeding up, because the interpreter releases that lock around blocking calls, which is why treating threads as useless is wrong as a general claim.
Build the lost update: two threads each incrementing a shared counter a hundred thousand times, and show a final value below the expected sum, because an increment is a load, an add and a store and the thread can be suspended between them. Fix it with a lock and then measure what the lock costs.

Deliverable: Three measurements, threads against processes on CPU work, threads on I/O work, and a demonstrated lost update, each with the mechanism written underneath.

Day 5 — Debugging as a procedure rather than an instinct

Work one real failure as a bisection: find a revision or an input size where it is good and one where it is bad, halve repeatedly, and state the two assumptions bisection needs, that the property changes exactly once across the range and that the test is reliable.
Minimise one failing input to the smallest version that still fails, and record how many rounds it took.
Keep a hypothesis log for one bug in three columns, what I believe, what would disprove it, what I observed, and stop yourself the first time you are about to change two things at once.

Deliverable: One bug worked to root cause with a written hypothesis log and a minimised reproducing input.

Day 6 — Tests that catch the bug you are about to write

Implement an LRU cache with a capacity bound, then write the three test cases that would catch an off-by-one in eviction: insert exactly capacity items and assert nothing was evicted, insert one more and assert the least recently used key is the one gone, and read an old key just before that insert so the eviction victim changes.
Add a property test comparing your implementation against a deliberately slow reference, an ordered list scanned linearly, over a few thousand random operation sequences, because a slow reference finds the cases you would not have thought to write.
Write one numeric test that fails under exact equality and passes with a tolerance, and state why the tolerance has to be relative rather than absolute once the magnitudes grow.

Deliverable: An LRU implementation with three boundary tests, one property test against a slow reference, and one tolerance-based numeric test.

Day 7 — Debug something broken, out loud

Have someone plant three defects in a two-hundred-line program, an off-by-one, a shared mutable state bug, and a wrong error-handling path, then find them while narrating, under a fixed rule: state the hypothesis before touching anything.
Time each one and record which tool found it, reading, a printed value, a debugger, or a test, because the question asked in interviews is how you would find it rather than what it was.
Write the sentence you will use when you do not yet know the cause, one that names the next measurement instead of offering a guess.

Deliverable: A recorded debugging session with time-to-find per defect and the method that found each.

Disagreement needs an outcome you did not choose

Conflict answers where you were right and everyone came round are the weakest ones. Stronger: the evidence you went and collected, what would have changed your mind, and what you did in the weeks after the call went against you. Implementing a design you argued against, properly, is a specific and checkable behaviour.

How do you handle class imbalance in medical imaging datasets?

Medium · behavioural and collaboration

Approach

State the situation in two sentences and spend the rest on the reasoning.
Give the blast radius: what could have broken, and what you measured.
Pick a story where you made the decision, not one where you watched it.

Follow-up

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

Ship under a deadline and bound the debt you chose

Medium · pagination · technical debt · tradeoffs

You have four days to ship a tenant-facing listing endpoint. The version you would defend uses keyset pagination over (tenant_id, status, updated_at DESC, resource_id DESC); the version you can finish uses LIMIT/OFFSET with no matching index. Describe a deadline call you actually made of this shape: what you shipped, what you knowingly deferred, how you bounded the damage with a mechanism rather than an intention, and the specific numeric condition that would force the follow-up. Name who you told and where you wrote it down.

Approach

Name the deferred failure precisely instead of calling it slow. OFFSET n makes the database produce and discard n rows, so cost grows with page depth; without an index matching the sort, every matching row is read and sorted before the limit applies; and rows inserted between two page fetches shift across the boundary so items are skipped or repeated with nothing in the response to signal it.
Bound the blast radius with something mechanical rather than a promise: cap maximum page depth, cap page size, restrict the endpoint to one internal caller, or keep it behind a flag. State which failure each cap removes and which it leaves standing.
Attach a number to the trigger and wire it to an alarm: the first tenant crossing N resources, or the endpoint's p99 crossing its share of the 400 ms budget, so the debt announces itself instead of waiting to be remembered.
Write it where the next engineer looks, which is the code and the ticket, not a chat message: what was deferred, why, the cap, and the trigger.
Report what actually happened in your real example, including the case where the trigger never fired and the debt was correctly never repaid.

Follow-up

At what page depth does the offset version breach your latency budget, given your page size and row counts?
What breaks first when you switch to keyset pagination later, and what does a client holding an old page token see?
Who would have overruled you if you had asked for two more days, and did you ask?

Estimate work you have never done and defend the range

Hard · estimation · backfills · expand-contract

You are asked to estimate a change you have never attempted: add a column to a 100-million-row table, populate it, move reads across, and drop the old shape. Give a range with the assumptions that generate it, including batch size, the signal your backfill throttles on, and wall-clock hours, and name the three unknowns that would move the number most. Then describe a real estimate you gave under comparable ignorance: how you expressed its uncertainty, what you committed to, and how wrong you turned out to be.

Approach

Decompose into independently deployable steps before estimating anything: add the column nullable, write both shapes, backfill in batches, verify, move reads, stop writing the old shape, drop it. That is four deploys spread over days, and the calendar estimate is dominated by them rather than by the loop's runtime.
Do the arithmetic aloud for the part that has arithmetic in it: batch size times number of batches times per-batch duration, at a write rate the primary can absorb alongside roughly 1.2k writes per second of production traffic. The loop is throttled by replication lag and lock waits, not by how fast it can issue statements.
Price the schema step by its lock rather than its statement duration. In PostgreSQL an ALTER TABLE taking ACCESS EXCLUSIVE waits for every open transaction on that table while later queries queue behind it, so a millisecond change issued during a thirty-second analytics query stalls that table for thirty seconds. Adding a nullable column with a non-volatile default avoids a rewrite from version 11; a new index wants CREATE INDEX CONCURRENTLY, which cannot run inside a transaction block and leaves an invalid index behind if it fails.
Express the answer as a range whose endpoints each trace to a stated assumption, then name the cheapest experiment that collapses it, which is almost always running one real batch against the real table and multiplying.
Commit to a checkpoint rather than a completion date: the day you report a measured number from that first batch. That is a promise you can keep under uncertainty, and it is what the asker actually needs in order to plan.

Follow-up

How do you verify the backfill genuinely finished, given rows written by production traffic while it ran?
Where does the backfill resume from after a worker is killed mid-batch, and what makes that resume point trustworthy?
Your first batch comes back ten times slower than assumed. What do you tell the person waiting on the estimate, and when?