AstraZeneca Machine Learning Engineer Interview Questions & Guide 2026

The 60-second brief

As a Machine Learning Engineer at AstraZeneca, you occupy a pivotal intersection between cutting-edge data science and life-saving pharmaceutical innovation. You are not merely building models; you are architecting the analytical engines that accelerate drug discovery, optimize clinical trial design, and refine supply chain logistics. Your work directly impacts the speed at which transformative therapies reach patients globally.

The loop does not sample the job evenly, and arguing about that in the room costs you. Daily work is mostly incremental change inside code someone else wrote, while the loop samples narrow slices of it; prepare for the slices and save the realism argument for your questions at the end.

Focus areas

• Trace a symptom to a mechanism under load
• Make every write idempotent under retry
• Paginate large result sets with keyset cursors

Overview

This role requires a high degree of technical rigor and a passion for applying advanced AI to complex biological and operational datasets. You will collaborate with cross-functional teams, including biologists, clinicians, and data scientists, to translate ambiguous business challenges into scalable, production-ready machine learning solutions. Success in this role demands the ability to navigate the complexities of a highly regulated industry while maintaining the agility of a modern engineering organization.

The AstraZeneca interview process

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

Initial Screening

The title covers product work, platform work, infrastructure, mobile and frontend, and those are different jobs with different loops behind them. A screening call is the cheapest place to find out which one the seat is, and asking reads as experienced rather than fussy. The questions that separate them: what the team is on call for, what the last three projects were, and whether any round happens inside an existing repository instead of a blank file. Then say which of that you have done and which you have not. Claiming the whole posting is the fastest way to be found out one round later.

What to demonstrate

• Whether you can locate your experience inside one flavour of the role honestly instead of claiming the entire requirements list
• Whether you name what you have not done, which an experienced screener reads as a level signal and can plan the loop around
• Whether what you want next matches what the seat is: someone who wants greenfield work landing on a team that mostly operates an existing system is a hire that leaves within the year

How to prepare

• Mark every line of the posting as done, adjacent or new, and write one sentence for each adjacent line naming the closest thing you actually built
• Split your last two years into rough percentages across feature work, operating and debugging live systems, and design or review, so a question about scope gets numbers rather than adjectives
• Bring three questions that discriminate between seats: what the team is paged for, how much of the work is changing existing code versus standing up something new, and what shipped in the last quarter

Technical Assessments

What this round decides is narrow: whether you can produce code that runs and is correct on inputs nobody showed you. An elegant solution that does not compile scores below a plain one that does, so write a correct brute force first, say out loud that you know its cost, and improve it with the working version still on screen. What separates strong answers is who finds the broken case. Trace your own code against an empty input, a single element, and duplicate keys before you say you are finished, because being told is far more expensive than noticing.

What to demonstrate

• Whether degenerate inputs get checked without being asked for: an empty collection, one element, every element equal, and the extreme value the input type allows
• Whether the complexity you state matches the code you actually wrote, including a sort or a copy sitting inside a loop
• Whether the finished answer is verified against the worked examples before you call it done, rather than assumed correct because the code reads correctly

How to prepare

• Take five problems you have already solved and, without running anything, write down what each returns for empty input, a single element, and all-duplicates. Then run them and count how many you predicted wrong.
• Drill the brute force as its own skill: on ten problems, write only the obviously-correct slow version and time how long it takes to get it passing. If that is more than a few minutes, that is what to practise, not the optimal version.
• Add a fixed last step before you submit anything, reading only the loop bounds and the initial value of each accumulator, which is where most off-by-one errors live

Presentation Round

An unlabelled round is first an information problem, and the cheapest information is free. Whoever schedules it can usually tell you how long it runs, who will be in the room and what they work on, whether you will be writing code and in what environment, and whether anything is being sent beforehand. Ask in writing so the answer is on record, then prepare for the two or three formats those answers still leave open instead of betting on one. What separates a strong candidate is not guessing right; it is having an opening that works whichever one it turns out to be.

What to demonstrate

• Whether you can start work from an ambiguous brief, since tolerating a vague scope without stalling is the same thing the job asks for
• Whether the questions you asked beforehand were ones that change your preparation, such as duration, medium and who is joining, rather than ones whose answers you could not have acted on
• Whether you adapt when the round turns out to be something other than what you were told, instead of spending the first ten minutes visibly recalibrating

How to prepare

• Send one short scheduling message asking four things: how long, who is joining and what they work on, whether you will be writing code and where, and whether to prepare anything in advance. Treat a vague reply as real information, since it means the round is loosely structured and you will be shaping it yourself.
• Write one opening that works in any of the formats still open: restate in your own words what you have been asked to do, then ask which of two directions is more useful to them. Say it aloud until it stops sounding recited.
• Set up for the two most likely formats before the call starts, with a blank editor in the language you would choose and a shared document you can type into, so a format surprise costs you nothing in the first minutes

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.

Shipping a migration and the code that depends on it as a single change

During any rolling deploy, and for as long as a rollback remains possible, old and new code execute against the same schema at the same time. A migration that drops or renames a column breaks every instance that has not restarted yet, and code that requires a column the migration has not applied breaks every instance that restarted early. The discipline is expand then contract: add the new column nullable, write both shapes, backfill in batches, move reads across once the backfill is verified, and only then stop writing the old shape and drop it - four deploys, usually spread over days. It feels disproportionate until the first rollback, at which point it is the only reason the previous version still runs.

Going silent while thinking

Narrate the candidates and why you are discarding them, even in fragments: sorting first would make this a two-pointer scan, but it destroys the original indices, which the output needs. From the other side of the table, a candidate thinking hard and a candidate stuck are indistinguishable until one of them speaks.

Hardcoding to the sample inputs

Solve the stated problem rather than the two examples; special-casing a literal to make a sample pass is obvious immediately and reads as either a misunderstanding or an attempt to fake progress. If you genuinely cannot generalise yet, say which part is a stub and what would replace it.

Questions & practice

Walk us through a machine learning project where you encountered a significant roadblock and how you resolved it.

Medium · machine learning fundamentals

Approach

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

Follow-up

• How would you know the model is overfitting?
• Where could label leakage enter this setup?

How do you approach feature engineering for high-dimensional biological data?

Medium · machine learning fundamentals

Approach

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

Follow-up

• Where could label leakage enter this setup?
• What changes if the classes are heavily imbalanced?

Explain the trade-offs between different model architectures for [specific data type].

Medium · machine learning fundamentals

Approach

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

Follow-up

• What changes if the classes are heavily imbalanced?
• Where could label leakage enter this setup?

What methods do you use to ensure model interpretability in highly regulated projects?

Medium · machine learning fundamentals

Approach

1. Name the simplest model that could work and what would make you move past it.
2. State the learning problem: the label, the unit of prediction and how the model is used.
3. Say how you would validate it, and where leakage could enter the split.

Follow-up

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

Merge partitioned event streams into one ordered feed with bounded lateness

Hard · k-way merge · watermarks · out-of-order streams

The read-model service consumes 64 log partitions carrying about 4,000 events per second in total. Each partition is ordered within itself, but partitions drift by up to 30 seconds, and the activity feed must present a tenant's events in occurred_at order. Produce the merge. State its complexity, the buffer it requires in events and in bytes, what happens when one partition is idle, and what you do with an event that arrives after you have already emitted its position. Payloads average 1 KB.

Approach

1. Merge with a min-heap over the 64 partition heads keyed on (occurred_at, event_id): O(log P) per event and O(n log P) overall. The tie-break on event_id is what makes the output deterministic when two partitions carry the same millisecond, which matters because the feed is paginated and a non-deterministic order reorders pages under the reader.
2. Emitting the heap head is only correct once every partition has produced everything up to that timestamp, so the emit condition is a watermark: the minimum across partitions of the highest occurred_at seen, less the allowed lateness. Events are held until the watermark passes them, which is what turns individually ordered streams into a jointly ordered one.
3. Size the buffer from the lateness rather than guessing: 4,000 events per second times 30 seconds is 120,000 buffered events, and at 1 KB each about 120 MB of heap. That number is the real price of the ordering guarantee and belongs in front of whoever asked for it.
4. Handle the idle partition explicitly, because it fails the feed rather than corrupting it: a partition with no traffic never advances its own maximum, so the watermark freezes and output stops entirely. Either every partition emits a periodic idle marker carrying the broker's current time, or the watermark falls back to wall clock for a partition silent beyond a threshold.
5. Choose the late-event policy from what the projection is keyed on. The projection upserts on (aggregate_id, aggregate_version) and discards a version it has already applied, so a late event is safe to apply out of order and correctness never depended on the merge at all. Apply it, recompute the affected feed page, and count lateness so the 30-second budget can be re-derived from data rather than folklore.
6. Say what the merge does not buy: ordering is guaranteed within one aggregate by the log's partitioning, and no watermark makes the cross-aggregate order authoritative. Two events from different aggregates in the same millisecond have no true order, so the feed's order is a presentation choice that must be stable rather than correct.

Worked solution (35 min)

1. Write the heap comparator on (occurred_at, event_id) and the per-partition head refill.
2. Write the watermark computation and the emit-loop condition, then list which buffered events are held at a chosen instant.
3. Compute the buffer at 4,000 events per second, 30 seconds and 1 KB per event, and state what fraction of a worker's heap that represents.
4. Add the idle-partition marker and trace the watermark with one silent partition, both with and without the marker.
5. Write the late-event path and name the key that makes applying it safe.

Expected result: A 64-way min-heap merge at O(n log P) with a deterministic (occurred_at, event_id) comparator, a watermark of the per-partition minimum less 30 seconds gating emission, a stated buffer of 120,000 events and roughly 120 MB, idle markers so a silent partition cannot freeze the watermark, and a late-event policy justified by the projection's idempotency on (aggregate_id, aggregate_version).

Verify it yourself

• Feeding the partitions in an adversarial interleaving produces the same output sequence as feeding them in timestamp order.
• Silencing one partition stalls output without the idle marker and does not stall it with the marker.
• Applying a late event twice leaves the projection byte-identical, which is what permits the out-of-order path to exist.

Follow-up

• The lateness budget is raised to five minutes. What is the new buffer, and what besides memory changes?
• The consumer restarts. Where does it resume from, and what does the feed look like for the first 30 seconds?
• One partition is ten minutes behind because its producer is slow. Do you stall the feed or emit without it?

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

1. 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.
2. 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.
3. 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.
4. 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.
5. 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.
6. 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.

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?

Stop tag and share joins from fanning out a page

Medium · join fan-out · semi-joins · many-to-many

resource_tag is (resource_id, tag_id) with PK (resource_id, tag_id); resource_share is (resource_id, shared_with_user_id, permission). The tagged-and-shared listing inner-joins resource to both, filters tenant_id, tag_id = ANY($2) and shared_with_user_id = $3, orders by updated_at DESC and takes 50. Pages come back with fewer than 50 distinct resources and the total in the header is far too high. Explain the row multiplication, rewrite both the page query and the count query so each is correct, and name the index each one needs. PostgreSQL 16.

Approach

1. Do the arithmetic against the predicates that are actually there. An inner join emits one row per matching child row, and both joins are filtered: tag_id = ANY($2) admits only the requested tags, shared_with_user_id = $3 admits one user's share rows. So a resource holding three of the requested tags and shared with $3 once yields three rows, not one — the multiplier is its count of matching tags times its share rows for that single user, and that second factor is 1 unless the table admits duplicate (resource_id, shared_with_user_id) pairs. LIMIT 50 then limits rows rather than resources, and COUNT(*) counts pairs — the header is the product, not the population.
2. Reject DISTINCT as the fix. It deduplicates after the product has been built, so the planner must materialise and sort the fanned-out set before the LIMIT can apply, and it leaves any SUM or AVG in the same select list wrong.
3. Rewrite both filters as semi-joins, keeping resource as the only row source: AND EXISTS (SELECT 1 FROM resource_tag rt WHERE rt.resource_id = r.resource_id AND rt.tag_id = ANY($2)) and the same shape against resource_share. A semi-join stops at the first match per resource and preserves the driving index order, so ORDER BY updated_at DESC, resource_id DESC LIMIT 50 still stops after 50 rows.
4. Count with the same predicates and no join at all: SELECT count(*) FROM resource r WHERE r.tenant_id = $1 AND r.status = 'active' AND EXISTS (...) AND EXISTS (...). Nothing multiplies a resource, so the number is the population.
5. Attach the tags for display after the page has been cut — LEFT JOIN LATERAL (SELECT array_agg(rt.tag_id) FROM resource_tag rt WHERE rt.resource_id = p.resource_id) ON TRUE over the 50 returned rows. Aggregate over the page, never over the tenant.
6. Index both directions and say which query each serves: PK (resource_id, tag_id) serves the lateral lookup, (tag_id, resource_id) serves the EXISTS probe by tag, and resource_share needs (shared_with_user_id, resource_id) for the same reason. An index covering one direction only leaves the other as a scan.

Worked solution (30 min)

1. Build a tenant where each resource carries 0-5 tags from a 20-tag vocabulary and is shared with 0-4 distinct users, then bind $2 to three tags and $3 to a user holding shares on about half the resources. Run the joined query and compare its row count to the distinct resource count on page one.
2. Run COUNT(*) on the joined shape and on the EXISTS shape and compare both to a ground truth computed from distinct ids; then give $3 a second permission row on 10% of resources and record which of the two counts moves.
3. EXPLAIN both page queries and compare rows-read plus the presence of a Sort or HashAggregate node above the join.
4. Add (tag_id, resource_id), re-run the EXISTS probe, and record the plan change on the inner side.

Expected result: The EXISTS page returns exactly 50 distinct resources and reads about 50 driving rows with no aggregate above the join. The EXISTS count matches the ground truth. The joined COUNT(*) is larger by roughly the average number of matching tags per qualifying resource — the tags it holds from $2, not its total tag count — times its share rows for $3, which is 1 until the duplicate permission rows exist and rises with them afterwards. The EXISTS count is unchanged by those duplicates.

Verify it yourself

• The page's resource_id list has no duplicates and is 50 long whenever at least 50 rows match.
• count(*) from the rewrite equals count(DISTINCT r.resource_id) over the joined shape.
• EXPLAIN shows a semi-join node on the inner side rather than a full join followed by a unique or aggregate step.

Follow-up

• The filter changes from 'any of these tags' to 'all of these tags'. Rewrite it and state what it costs relative to the ANY form.
• A resource can be shared with the same user twice under different permissions. Does your count change, and should it?
• Where does the correct total come from when the tenant holds 4M resources and the header must not cost 200 ms?

How do you ensure your code and documentation are scalable and maintainable for team members?

Medium · system design

Approach

1. Fix the scope first: who calls this, how often, and what they do when it fails.
2. State the consistency you need, and where you are willing to be stale.
3. Name the read and write paths separately; they rarely have the same bottleneck.

Follow-up

• What breaks first when traffic grows ten times?
• How does this behave when that dependency is down for an hour?

How do you validate the performance of your models before they are used in decision-making processes?

Medium

Approach

1. Say what you would check first and why it is the highest-information step.
2. Clarify what is being asked and what a complete answer contains.
3. Work from the requirement backwards to the design.

Follow-up

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

Keep one tenant's bulk export from starving projection updates

Medium · backpressure · fair scheduling · head-of-line blocking · leases

The worker fleet runs about 600 jobs/second out of job_run, claimed under a lease and heartbeated while running. Most handlers finish under 200 ms; exports run for minutes. One tenant enqueues 50,000 exports. Design the scheduling so that tenant's exports slow down and nothing else does: the queue split, the concurrency caps and where they are enforced, the claim query, and the backpressure signal that stops accepting new work. State the in-flight worker count each class needs at its arrival rate, and what the cap costs the tenant that hits it.

Approach

1. Do the concurrency arithmetic first, because it determines the pool shapes. Required in-flight work is arrival rate times service time: 600/second of 200 ms handlers needs about 120 workers, while one export per second at 180 seconds needs 180 on its own. The long tail dominates any pool it shares, so a single pool sized from the mean is consumed by exports while 50 ms projections queue behind them. That is head-of-line blocking, and adding capacity does not fix it because the ratio is what is wrong.
2. Split by job class rather than by priority. Priority inside one pool still lets a running export hold its worker for the next four minutes - there is no preemption for a handler already executing. Separate queues and separate worker processes give each class a floor the other cannot take.
3. Enforce the per-tenant cap inside the claim, not at enqueue. Claim with SELECT ... WHERE status='queued' AND run_after <= now() AND job_type=$1 ORDER BY run_after, job_run_id FOR UPDATE SKIP LOCKED LIMIT 1, restricted to tenants currently under their cap. Derive the running count from rows in status 'running' with a live lease rather than from a counter incremented at claim and decremented at completion: the derived count self-heals when a worker dies, and the counter leaks a slot every time one does.
4. Make the ordering fair over tenants instead of over jobs. Strict FIFO with a cap still forces the claim to scan past 50,000 rows belonging to a capped tenant before it finds anyone else's work, so cost grows with backlog depth. Pick the tenant first from a small per-tenant queue-depth summary, then claim within that tenant, which makes the scan proportional to the number of active tenants instead.
5. Close the loop with backpressure and deduplication. Reject or defer enqueues once a class's depth passes a bound, and rely on UNIQUE (job_type, dedupe_key) WHERE status IN ('queued','running') so a retrying producer collapses into one row instead of multiplying the backlog. Watch oldest-queued-age per class rather than depth: depth means nothing without a service rate, while age is the user-visible latency directly.

Worked solution (25 min)

1. Compute required in-flight workers per class from arrival rate times service time, and show what one shared pool does when both classes contend.
2. Write the claim statement with the per-tenant restriction, and state how the running count is derived so it survives a worker being killed.
3. Trace what the claim scans when one tenant holds 50,000 queued rows, under FIFO and under tenant-first selection.
4. Choose the backpressure threshold and the signal it watches, then state what the producer receives when it trips.

Expected result: Separate queues and worker pools per job class, sized from rate times service time (about 120 for 200 ms work, about 180 for exports). A per-(job_type, tenant) cap enforced inside the claim with FOR UPDATE SKIP LOCKED, with the running count derived from live leases so it self-heals. Tenant-first selection keeps claim cost proportional to active tenants rather than backlog depth. Enqueue is rejected on class depth, duplicates collapse on dedupe_key, and the alarm is oldest-queued-age per class.

Verify it yourself

• A tenant with 50,000 queued exports never occupies more than its cap's worth of workers.
• Claim cost does not grow as that tenant's backlog grows.
• Killing a worker mid-job returns its cap slot within the lease interval with no manual step.
• Projection p99 is unchanged while the export backlog drains.

Follow-up

• An export outruns its lease, so two copies run. What must the handler do to make the second copy harmless, and what does it write to guarantee that?
• With a cap of 4, the tenant's 50,000 exports now take hours. What do you tell them, and does the answer change if they are the largest tenant on the platform?
• Queue depth is flat but oldest-queued-age is climbing. Name the mechanisms that produce that exact pair of signals.

Read latency spikes on a sixty-second sawtooth

Medium · cache stampede · single-flight · ttl · tail latency

The cached listing read path serves about 14k reads/second at an 85% hit rate. p99 sits at 35 ms for 57 seconds, jumps to 900 ms for 3, and repeats. During each spike the primary shows several hundred identical listing queries starting within the same millisecond, all carrying one large tenant's id. Cache entries use a 60-second TTL. Give the mechanism, the ordered checks, the fix, and the correctness hazard your fix must not introduce.

Approach

1. Match the period to a configured number before theorising about load. A spike every 60 seconds against a 60-second TTL is an entry expiring, and you confirm it by correlating spike timestamps with the entry's write time rather than with the traffic curve. If the period had matched a cron or a GC interval instead, this is a different investigation.
2. Establish the concurrency of the miss. Several hundred identical queries in one millisecond means the miss path has no coalescing: every request that arrives between expiry and repopulation recomputes. The herd size is that key's arrival rate times its recompute time, so at 1.2k reads/second for the hot key and a 250 ms recompute you expect about 300 concurrent misses, which matches what is observed.
3. Add single-flight on the miss path so one caller per key recomputes under a short-lived lock while the rest wait for its result. Prefer stale-while-revalidate where the read tolerates it: return the expired value immediately and refresh asynchronously, which removes the latency spike rather than serialising it into a queue of waiters.
4. De-synchronise the keys. Write TTLs with jitter, for example 60 seconds plus or minus 10%, so a deploy or a mass invalidation does not align every key on the same second and turn a per-key herd into a fleet-wide one.
5. Name the hazard the fix must not introduce. Serving a stale listing is acceptable only because the API reports the projection watermark, and a reader that loaded the old value before a write can repopulate the entry after the invalidation, so the bounded TTL is what actually caps staleness rather than the delete. Keep read-after-write pinned to the primary for the writing session regardless.
6. Verify on miss concurrency, not hit rate. The hit rate barely moves, because the herd is one miss multiplied; the number that must change is distinct origin queries per key per minute.

Follow-up

• The same sawtooth appears on a key that is invalidated on write rather than expired. Is that the same bug?
• How does your answer change if the recompute takes 4 seconds instead of 250 ms?
• What exactly does a client see during a stale-while-revalidate window, and how does the watermark let them tell?

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.

Bounding a vague ticket before writing code

When the requirements were thin, the interesting part is how you fenced the problem off: the assumption you wrote down, who you got to confirm it, the narrow version you shipped first so the rest stayed cheap to change. Guessing and being right is luck. Guessing in writing, where someone could correct you, is method.

Describe a situation where you had to influence others without having direct authority.

Medium · behavioural and collaboration

Approach

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

Follow-up

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

Tell us about a time you worked with a cross-functional team to deliver a project.

Medium · behavioural and collaboration

Approach

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

Follow-up

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

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

1. 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.
2. 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.
3. 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.
4. 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.
5. 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?