At Thinking Machines, a Software Engineer is more than just a developer who writes code. You are a systems builder, a problem solver, and a strategic partner who translates complex data science and machine learning capabilities into robust, production-grade enterprise software. The engineering team sits at the intersection of big data, artificial intelligence, and cloud infrastructure, building custom data platforms and scalable APIs that empower organizations to make data-driven decisions. The systems you build directly impact global enterprises, non-profits, and government agencies. Whether you are optimizing a high-throughput geospatial data pipeline, deploying large language models to production, or designing secure, resilient backend architectures, your work ensures that data is accessible, reliable, and actionable. Because Thinking Machines operates as a highly collaborative technology consultancy, engineers must possess both technical depth and the ability to understand the broader business outcomes of their technical decisions. This role is highly critical because of the sheer scale and complexity of the data systems involved. You will be expected to design systems that are not only performant but also elegant and maintainable. If you thrive in an environment that values autonomy, continuous learning, and cross-functional collaboration between data scientists, product managers, and clients, you will find this position both challenging and deeply rewarding.
Technical Screening
reportedInitial assessment which may involve an online test or a take-home backend coding exam.
What to demonstrate
- Initial assessment which may involve an online test or a take-home backend coding exam
- Depth in Data Structures (DSA)
How to prepare
- Answer aloud and timed: How would you structure a backend service to handle high-throughput, concurrent file uploads while maintaining data integrity?
- Answer aloud and timed: Explain your choice of database (SQL vs. NoSQL) for a system that requires fast read access but has highly unpredictable write spikes.
Take-Home Exam
reportedComplete a functional backend service in your chosen language, along with written technical and behavioral responses.
What to demonstrate
- Complete a functional backend service in your chosen language, along with written technical and behavioral responses
- Depth in Data Structures (DSA)
How to prepare
- Answer aloud and timed: How do you handle error propagation and logging in a distributed microservices architecture?
- Answer aloud and timed: Walk me through the process of optimizing a slow API endpoint that aggregates data from multiple external services.
Presentation and Panel Interview
reportedPresent your take-home solution to a panel, discussing code, architectural choices, and scaling strategies.
What to demonstrate
- Present your take-home solution to a panel, discussing code, architectural choices, and scaling strategies
- Depth in Data Structures (DSA)
How to prepare
- Answer aloud and timed: What strategies would you use to secure an API that handles sensitive geospatial or financial data?
- Answer aloud and timed: Why did you choose this specific framework or language for your coding exam, and what are the trade-offs of that choice?
Culture Fit Interview
reportedFinal interview to assess mutual alignment with the company's culture.
What to demonstrate
- Final interview to assess mutual alignment with the company's culture
- Depth in Data Structures (DSA)
How to prepare
- Prepare three examples from your own work, each with a decision you made and an outcome you can quantify.
- Re-read the description of the culture fit interview above and write down what you would ask to confirm before it.
PracHub editorial advice for the preparation topics above.
Structure your presentation clearly
When preparing your slides for the panel interview, start with the business problem, move to the high-level architecture, and then dive into the specific technical implementation details.
Test your take-home code thoroughly
Ensure your code runs out of the box. Provide clear setup instructions in your README. If an interviewer cannot run your code easily, it creates a poor first impression.
Be honest about trade-offs
No system is perfect. During your presentation, proactively point out the limitations of your design and explain how you would improve it if you had more time.
Focus on readability
Write clean, self-documenting code with sensible variable names and modular structures. Avoid overly clever or obfuscated code that is difficult to explain during a review.
Choose a category, try a prompt, then open its approach, worked solution or follow-up when you need it.
Archive a resource graph without breaking live references or recursing
Resources reference other resources within a tenant; for the largest tenant the reference table holds up to 2,000,000 nodes and 8,000,000 edges. Archiving a resource must archive everything reachable from it that nothing outside the set still references, refuse when a live external referrer exists, and terminate when references form cycles, which they legitimately do. Produce the archive order and the refusal list, targeting O(V+E). Say what stops the traversal crossing a tenant boundary, and why recursion is the wrong control structure at this size.
Approach
- Load the subgraph with the tenant predicate on both endpoints of the edge, not only on the side you started from. Scoping the left table alone is the classic cross-tenant leak: one mis-entered edge then pulls another tenant's resources into the traversal and, worse, into the archive.
- Traverse iteratively with an explicit stack. A 2,000,000-node graph can hold a chain deep enough to exhaust a native stack in the low tens of thousands of frames, and that failure is a process crash rather than an error you can return.
- Treat cycles as data rather than corruption: compute strongly connected components with Tarjan in O(V+E) using its own explicit stack, then condense. The condensation is a DAG, so a topological order over it gives the archive order, and every member of a component archives in one transaction because no order within a cycle is valid.
- Decide refusals with reverse edges. A candidate is archivable only if every in-edge originates inside the candidate set, so build the transpose or count in-degrees restricted to the visited set, and emit each blocked resource with the id of the external referrer, which is the only part of the answer an operator can act on.
Follow-up
- The graph is read in one query and the archive writes a minute later. What can change in between, and how do you make the write safe?
- The candidate set is 400,000 resources. Is that one transaction, and if not, what does a half-finished archive look like to a reader?
Merge partitioned event streams into one ordered feed with bounded lateness
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
- 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.
- 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.
- 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.
- 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.
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?
Track a rolling failure rate per destination for circuit decisions
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.
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?
Explain your choice of database (SQL vs. NoSQL) for a system that requires fast read access but has highly unp
Explain your choice of database (SQL vs. NoSQL) for a system that requires fast read access but has highly unpredictable write spikes.
Approach
- Name the grain you start from and join outward from it.
- Check whether any join is one-to-many before aggregating, or the sums inflate.
- Say which index the query would use, and what makes it unusable.
- Handle the rows that do not match: that is usually the actual question.
Follow-up
- How does the query change if that join becomes one-to-many?
- What happens to this when the table is ten times larger?
Denormalise tenant onto revisions and backfill it live
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.
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?
How would you structure a backend service to handle high-throughput, concurrent file uploads while maintaining
How would you structure a backend service to handle high-throughput, concurrent file uploads while maintaining data integrity?
Approach
- Fix the scope first: who calls this, how often, and what they do when it fails.
- Name the read and write paths separately; they rarely have the same bottleneck.
- Choose a partition key and say what query it makes expensive.
- State the consistency you need, and where you are willing to be stale.
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 handle error propagation and logging in a distributed microservices architecture?
How do you handle error propagation and logging in a distributed microservices architecture?
Approach
- Fix the scope first: who calls this, how often, and what they do when it fails.
- Name the read and write paths separately; they rarely have the same bottleneck.
- Choose a partition key and say what query it makes expensive.
- State the consistency you need, and where you are willing to be stale.
Follow-up
- What breaks first when traffic grows ten times?
- How does this behave when that dependency is down for an hour?
Walk me through the process of optimizing a slow API endpoint that aggregates data from multiple external serv
Walk me through the process of optimizing a slow API endpoint that aggregates data from multiple external services.
Approach
- Say who the caller is and what they do when the call fails halfway.
- Define the identity of a request so a retry cannot double-apply it.
- Separate accepted, pending, failed and confirmed; they are different facts.
- Design the error taxonomy before the success shape; callers branch on it.
Follow-up
- What happens if the caller retries after a timeout?
- How does a client discover it is on an old version of this contract?
What strategies would you use to secure an API that handles sensitive geospatial or financial data?
What strategies would you use to secure an API that handles sensitive geospatial or financial data?
Approach
- Say who the caller is and what they do when the call fails halfway.
- Define the identity of a request so a retry cannot double-apply it.
- Separate accepted, pending, failed and confirmed; they are different facts.
- Design the error taxonomy before the success shape; callers branch on it.
Follow-up
- What happens if the caller retries after a timeout?
- How does a client discover it is on an old version of this contract?
Why did you choose this specific framework or language for your coding exam, and what are the trade-offs of th
Why did you choose this specific framework or language for your coding exam, and what are the trade-offs of that choice?
Approach
- Clarify what is being asked and what a complete answer contains.
- State your assumptions explicitly before working the problem.
- Say what you would check first and why it is the highest-information step.
- Work from the requirement backwards to the design.
Follow-up
- What assumption would you test first?
- How would you know your answer was wrong?
How does the architecture you designed for the take-home test support future business scaling or feature addit
How does the architecture you designed for the take-home test support future business scaling or feature additions?
Approach
- Fix the scope first: who calls this, how often, and what they do when it fails.
- Name the read and write paths separately; they rarely have the same bottleneck.
- Choose a partition key and say what query it makes expensive.
- State the consistency you need, and where you are willing to be stale.
Follow-up
- What breaks first when traffic grows ten times?
- How does this behave when that dependency is down for an hour?
If a client requested a 50% reduction in cloud infrastructure costs for this service, what architectural chang
If a client requested a 50% reduction in cloud infrastructure costs for this service, what architectural changes would you propose first?
Approach
- Fix the scope first: who calls this, how often, and what they do when it fails.
- Name the read and write paths separately; they rarely have the same bottleneck.
- Choose a partition key and say what query it makes expensive.
- State the consistency you need, and where you are willing to be stale.
Follow-up
- What breaks first when traffic grows ten times?
- How does this behave when that dependency is down for an hour?
Explain how your system handles failure states to a business stakeholder who is worried about service downtime
Explain how your system handles failure states to a business stakeholder who is worried about service downtime.
Approach
- Clarify what is being asked and what a complete answer contains.
- State your assumptions explicitly before working the problem.
- Say what you would check first and why it is the highest-information step.
- Work from the requirement backwards to the design.
Follow-up
- What assumption would you test first?
- How would you know your answer was wrong?
p99 jumped on one listing filter while p50 stayed flat
After a release that added an owner_user_id filter to the resource listing, p99 rose from 90 ms to 1.9 s while p50 stayed at 40 ms. Traffic and row counts are unchanged. resource carries the index (tenant_id, status, updated_at DESC, resource_id DESC). The new query filters tenant_id and owner_user_id, orders by updated_at DESC, resource_id DESC, and takes 20 rows. On PostgreSQL, explain the shape of the regression, prove it from a query plan, and give the index you would add.
Approach
- Start from the shape. A flat p50 with a moved p99 means a subset of requests changed cost, not all of them, so the first job is naming the subset. Bucket the endpoint's latency by the tenant's row count; the natural hypothesis is that large tenants are a small share of requests and all of the tail.
- Get the plan for the new query on a large tenant with EXPLAIN (ANALYZE, BUFFERS). Expect an index scan over the tenant's range, a filter discarding most of it, then a Sort feeding the Limit, possibly reporting Sort Method: external merge Disk. Read actual rows on the scan node, not estimated.
- Explain why the existing index cannot serve it. A composite B-tree is seekable only as a left prefix, and with no equality predicate on status the scan cannot treat updated_at as an ordering, because rows in the tenant's range are ordered by status first. Everything matching must be read and sorted before LIMIT 20 can apply, so a tenant with 400,000 rows pays 400,000 rows to return 20.
- Add (tenant_id, owner_user_id, updated_at DESC, resource_id DESC). Equality on the first two columns leaves the index ordered by updated_at within that pair, so the plan becomes an index scan that stops after 20 rows with no Sort node. PostgreSQL can scan a B-tree backwards, so the DESC markers matter only if the two sort columns ever disagree in direction; keeping them explicit documents the order the keyset cursor depends on.
Follow-up
- The endpoint paginates with OFFSET. What does page 500 cost with your index, and what does the keyset version cost?
- How would you have caught this before release, given that a 10,000-row seed database produces the same plan shape at an unnoticeable cost?
Built from the rounds and topics Thinking Machines candidates report.
Prepare, practise & reflect
One practical outcome each day. Spend longer where you need it.
0 / 7 done01Map the Thinking Machines loop
- Write out the reported sequence: Technical Screening, Take-Home Exam, Presentation and Panel Interview, Culture Fit Interview.
- For each round, write one sentence on what it is judging, from the description above, and mark the one you are least ready for.
Deliverable: A one-page map of the 4 reported rounds, with the weakest marked.
02Work Data Structures (DSA)
- Spend the session on Data Structures (DSA), which Thinking Machines candidates report being tested on.
- Write one worked example in Data Structures (DSA) and time yourself on it.
Deliverable: One timed worked example in Data Structures (DSA).
03Work Algorithms (Coding Practices)
- Spend the session on Algorithms (Coding Practices), which Thinking Machines candidates report being tested on.
- Write one worked example in Algorithms (Coding Practices) and time yourself on it.
Deliverable: One timed worked example in Algorithms (Coding Practices).
04Work Backend Development (General)
- Spend the session on Backend Development (General), which Thinking Machines candidates report being tested on.
- Write one worked example in Backend Development (General) and time yourself on it.
Deliverable: One timed worked example in Backend Development (General).
05Answer out loud: Backend Engineering & System Design
- Answer aloud, timed: How would you structure a backend service to handle high-throughput, concurrent file uploads while maintaining data integrity?
- Answer aloud, timed: Explain your choice of database (SQL vs. NoSQL) for a system that requires fast read access but has highly unpredictable write spikes.
Deliverable: Spoken answers to 2 reported Backend Engineering & System Design question(s), under time.
06Answer out loud: Technical Presentation & Business Translation
- Answer aloud, timed: Why did you choose this specific framework or language for your coding exam, and what are the trade-offs of that choice?
- Answer aloud, timed: How does the architecture you designed for the take-home test support future business scaling or feature additions?
Deliverable: Spoken answers to 2 reported Technical Presentation & Business Translation question(s), under time.
07Answer out loud: Behavioral & Collaboration
- Answer aloud, timed: Tell me about a time you had to work with a data scientist or product manager who had a different technical perspective than you. How did you reach a consensus?
- Answer aloud, timed: Describe a situation where you had to quickly learn a new technology or framework to deliver a project. What was your approach?
Deliverable: Spoken answers to 2 reported Behavioral & Collaboration question(s), under time.
Expand any day for tasks and deliverables. Your progress is saved on this device.
Behavioural rounds judge the decision you made and what it cost.
Tell me about a time you had to work with a data scientist or product manager who had a different technical pe
Tell me about a time you had to work with a data scientist or product manager who had a different technical perspective than you. How did you reach a consensus?
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.
- Give the blast radius: what could have broken, and what you measured.
- Name the disagreement and how you resolved it with evidence.
Follow-up
- What would you do differently if you ran that again?
- How did you know your change caused the improvement?
Describe a situation where you had to quickly learn a new technology or framework to deliver a project. What w
Describe a situation where you had to quickly learn a new technology or framework to deliver a project. What was your approach?
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.
- Give the blast radius: what could have broken, and what you measured.
- Name the disagreement and how you resolved it with evidence.
Follow-up
- What would you do differently if you ran that again?
- How did you know your change caused the improvement?
How do you balance the pressure of meeting a tight client deadline with the desire to write clean, refactored
How do you balance the pressure of meeting a tight client deadline with the desire to write clean, refactored code?
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.
- Give the blast radius: what could have broken, and what you measured.
- Name the disagreement and how you resolved it with evidence.
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 project you worked on that failed or did not meet expectations. What did you learn from it?
Tell me about a project you worked on that failed or did not meet expectations. What did you learn from it?
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.
- Give the blast radius: what could have broken, and what you measured.
- Name the disagreement and how you resolved it with evidence.
Follow-up
- What would you do differently if you ran that again?
- How did you know your change caused the improvement?
- 01
Tell me about a time you had to work with a data scientist or product manager who had a different technical perspective than you. How did you reach a consensus?
- 02
Describe a situation where you had to quickly learn a new technology or framework to deliver a project. What was your approach?
- 03
How do you balance the pressure of meeting a tight client deadline with the desire to write clean, refactored code?
- 04
Tell me about a project you worked on that failed or did not meet expectations. What did you learn from it?
How difficult is the Software Engineer interview process at Thinking Machines?
Candidates generally describe the difficulty as average but highly practical. The process is rigorous because it tests a broad range of skills—including coding, system design, and presentation—rather than just algorithm memorization. Coming prepared to discuss and defend your code is key to success.
Thinking Machines Software Engineer candidate reports ↗What differentiates successful candidates in this process?
Successful candidates demonstrate a strong balance of technical depth and communication. They don't just write functional code; they can explain their design decisions, discuss architectural trade-offs, and clearly articulate how their technical choices deliver business value.
Thinking Machines Software Engineer candidate reports ↗What is the culture like within the engineering team?
The engineering culture at Thinking Machines is highly collaborative, transparent, and autonomous. Engineers are encouraged to take ownership of their projects, continuously learn new technologies, and work closely with cross-functional teams of data scientists and product managers.
Thinking Machines Software Engineer candidate reports ↗How long does the entire interview process typically take?
The process is known for being smooth and responsive. From the initial application or screening to the final offer, the timeline typically spans two to four weeks, depending on how quickly you complete the take-home assessment.
Thinking Machines Software Engineer candidate reports ↗What topics does Thinking Machines test in interviews?
Thinking Machines interviews most often cover Data Structures (DSA), Research Engineering, Infrastructure Engineering, Algorithms (Coding Practices), and Training Systems. The exact emphasis depends on the specific role you apply for.
Thinking Machines Software Engineer candidate reports ↗Sources & methodology 3 sources ↗
No official company page is cited. Rounds and questions come from candidate reports and PracHub editorial material; each source shows the date it was read.
- 01Thinking Machines Software Engineer candidate reports ↗
Company-reported rounds, questions and FAQ.
Candidate reports · Accessed 2026-09-22 - 02PracHub Software Engineer practice ↗
PracHub practice material, not company-reported.
PracHub page · Accessed 2026-09-22 - 03PracHub preparation framework ↗
PracHub preparation guidance.
PracHub page · Accessed 2026-09-22
Explore related engineering environments and compare the preparation focus.
Aerospike Software Engineer Interview Questions & Guide 2026
Aerospike database-engine preparation through an original single-owner Python key-value store with bounded logical index accounting, disk-offset reads, conditional writes, tombstones, process-restart recovery and 3,240 checks. Existing 14 p
Roku Software Engineer Interview Guide
Prepare for Roku software interviews with tested ring-buffer and LRU models, invariants, role-specific fundamentals and clear concurrency follow-ups.
Zelis Healthcare Software Engineer Interview Guide
What Zelis Healthcare asks Software Engineer candidates, from reported rounds and questions.
Wurl Software Engineer Interview Guide
What Wurl asks Software Engineer candidates, from reported rounds and questions.