Interview guides · Discovery Machine

Discovery Machine Software Engineer Interview Guide

Discovery Machine Software Engineer interview guide: reported rounds, questions and how candidates describe the process.

The 60-second brief

As a Software Engineer at Discovery Machine, you are at the intersection of advanced Artificial Intelligence (AI) research and mission-critical deployment. You will be responsible for transforming complex, applied research into scalable software solutions, specifically supporting the Department of Defense (DoD). Your work is not just about writing code; it is about building the intelligent entities, behavior models, and simulation environments that define the future of training and tactical readiness. This role is inherently collaborative and requires a high degree of technical versatility. You will work in a small, tight-knit team where your contributions directly influence the company’s vision.

This guide is scoped to a Software Engineer candidate at Discovery Machine.

Focus areas

Artificial Intelligence (AI) Engineering
Natural Language Processing (NLP)
Speech Recognition

The Discovery Machine interview process

Discovery Machine candidates report 5 rounds over 4-6 weeks. The stages below are what candidates describe, not a published process.

Online Application

Submit your application to express interest in the Software Engineer role.

What to demonstrate

Submit your application to express interest in the Software Engineer role
Depth in Artificial Intelligence (AI) Engineering

How to prepare

Answer aloud and timed: Can you walk me through your experience with Java, C#, or Python?
Answer aloud and timed: How do you approach designing a class structure for a complex simulation?

Phone Screen

Initial call to establish your technical baseline and interest in the role.

What to demonstrate

Initial call to establish your technical baseline and interest in the role
Depth in Artificial Intelligence (AI) Engineering

How to prepare

Be able to walk your CV end to end in two minutes, and say why this company specifically.
Have your salary expectations, notice period and location constraints ready, and ask for the rest of the loop in writing.

Technical Deep Dives

Engage in detailed technical discussions to assess your expertise.

What to demonstrate

Engage in detailed technical discussions to assess your expertise
Depth in Artificial Intelligence (AI) Engineering

How to prepare

Answer aloud and timed: Explain your understanding of AI behavior modeling.
Answer aloud and timed: Tell me about a time you had to explain a complex technical solution to a non-technical stakeholder.

Whiteboard Challenges

Participate in problem-solving exercises on a whiteboard to demonstrate your thought process.

What to demonstrate

Participate in problem-solving exercises on a whiteboard to demonstrate your thought process
Depth in Artificial Intelligence (AI) Engineering

How to prepare

Answer aloud and timed: How do you handle feedback during code or model reviews?
Answer aloud and timed: What attracts you to working in a small company environment like Discovery Machine?

Final Presentation

Present your past projects and experiences as part of the final assessment.

What to demonstrate

Present your past projects and experiences as part of the final assessment
Depth in Artificial Intelligence (AI) Engineering

How to prepare

Answer aloud and timed: Describe a time you had to mentor a team member or lead a design effort.
Answer aloud and timed: How do you stay updated with the latest advancements in AI and software engineering?

Where candidates lose points

Own your projects

If you list a project on your resume, be prepared to answer deep-dive questions about the architecture, the challenges faced, and the results.

Practice whiteboarding

Even in a casual conversation, be ready to sketch out a system design or a logic flow.

Research our domain

Familiarize yourself with the concept of Serious Games and how AI is used in training simulations.

Show your curiosity

Ask thoughtful questions about the projects the team is currently working on. It demonstrates genuine interest in our mission.

Questions & practice

Identify the heaviest tenants in a five-minute window under memory pressure

Medium · top-k · heavy hitters · streaming

The edge service handles about 3,000 requests per second across roughly 50,000 tenants, peaking near 9,000. Expose the 50 heaviest tenants by request count over the trailing five minutes so limits can be tightened before one tenant's backfill starves the fleet. You may not retain five minutes of raw records. Give the exact solution and its memory, then the bounded-memory approximation with its error stated as a formula, and say which you would ship and at what tenant cardinality that choice changes.

Approach

Do the exact version first, because it is affordable at this cardinality: a ring of 300 one-second counters per tenant, advanced lazily, is 1,200 bytes of counters per tenant and roughly 60 to 90 MB for 50,000 tenants with overhead. Carry a running total and subtract the bucket you overwrite so a window read is O(1) rather than 300 adds.
Extract the top 50 with a size-k min-heap over the tenant sums: O(d log k) for d tenants, against O(d log d) to sort them all. Maintaining the heap continuously instead of on query requires a tenant-to-heap-index map, because incrementing a count already inside the heap means sifting from a known position, and without that map you rebuild the heap on every request.
State the approximation precisely rather than gesturing at sketches. Misra-Gries with m counters retains every item whose true count exceeds N/(m+1), and each retained count underestimates the truth by at most N/(m+1). With m = 1,000 and N = 900,000 requests in the window the error is roughly 900 requests, which is fine for spotting a tenant sending 50,000 and useless for ranking two tenants 200 apart.
Say what breaks when the window slides: Misra-Gries and Space-Saving are insert-only and cannot be decremented as records age out. The workable construction is one summary per sub-window, say ten seconds, with 30 summaries merged at query time, and the merged error is the sum of the per-summary errors, so the bound degrades linearly in the number of sub-windows.

Follow-up

The heaviest tenant is heavy because of one export job rather than user traffic. Should the limiter treat those as the same tenant?
Two tenants sit tied at the boundary of the top 50. Does your answer flap, and does the flapping matter?

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.

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?

Collapse a redelivered event batch into per-aggregate high-water marks

Easy · hashing · at-least-once · aggregation

You drain a batch of up to 5,000,000 events, each (aggregate_id BIGINT, aggregate_version INT, event_type, payload). The log guarantees order within one aggregate only; the batch merges 64 partitions, and a relay failover has redelivered a range, so an older version for an aggregate can appear after a newer one. Given a map of last_applied_version per aggregate, produce the events worth applying, at most one per (aggregate_id, version), plus the count discarded. Target O(n) time. State the memory for 2,000,000 distinct aggregates and what you do when it does not fit.

Approach

One pass, one hash map from aggregate_id to the highest version kept, and a discard counter. An event whose version is at or below last_applied_version for its aggregate is dropped without further work, which is the whole reason the event carries its version rather than a delta. O(n) expected time, O(d) space in distinct aggregates.
Keep the maximum, never the last occurrence. The redelivered range means the final appearance of an aggregate in the batch can be an older version than one seen earlier in the same batch, so last-wins applies stale state over newer state and the projection regresses with no error anywhere.
Cost the memory instead of calling it large: an 8-byte key plus a 4-byte version is 12 bytes of payload, and an open-addressed table held at a 0.7 load factor costs roughly 17 bytes per entry before per-slot metadata, so 2,000,000 aggregates is tens of megabytes in a native layout and several times that in a runtime that boxes both key and value.
If the distinct set exceeds memory, partition on hash(aggregate_id) mod P and reduce each partition independently. Every event for one aggregate hashes to the same partition, so the per-partition result is exact and the merge is concatenation rather than a second reduction.

Follow-up

The payload is a patch rather than a snapshot, so applying only the highest version loses the intermediate changes. What changes in your reduction?
How do you detect that version 7 arrived while version 6 was never delivered, and what should the consumer do about the gap?

Explain why the owner filter ignores the listing index

Medium · composite index · query plans · sort elimination

The only index on resource is (tenant_id, status, updated_at DESC, resource_id DESC). A new endpoint returns one user's resources across all statuses, newest created first: WHERE tenant_id = $1 AND owner_user_id = $2 ORDER BY created_at DESC LIMIT 20. On a tenant with 2M rows it takes 900 ms and EXPLAIN shows a sort above a large scan. Explain precisely why the existing index cannot serve it, give the index that can, and state which of these the new index still will not help: owner_user_id alone across tenants; the same query ordered by updated_at. PostgreSQL 16.

Approach

Separate the two jobs an index does. For filtering, a composite btree is seekable only on a left prefix, so with no predicate on status the scan can at best range over tenant_id and test owner_user_id per row; PostgreSQL 16 has no btree skip scan to jump the unconstrained column.
For ordering, the index is sorted by (status, updated_at) within a tenant and not by created_at, so the LIMIT cannot stop early: every matching row is read and then sorted. That is the 'Sort Method: top-N heapsort' line, and it is why the plan reads 2M rows to answer with 20.
Derive the replacement from the access path — equality, equality, then the ordering column: CREATE INDEX CONCURRENTLY ON resource (tenant_id, owner_user_id, created_at DESC). The scan seeks to the (tenant, owner) range and walks 20 entries in order, so the Sort node disappears along with the row-read.
Treat INCLUDE (title, status) as conditional, not free. An index-only scan still visits the heap for any row whose page is not marked all-visible, so on a table taking 1.2k writes/second the win depends on autovacuum keeping the visibility map current, and the wider index costs more on every insert.

Follow-up

90% of rows are status='active'. Would a partial index WHERE status = 'active' change your answer, and for which of the three queries?
A dashboard runs this for 40 owners in one page load. What changes about the design?

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.

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?

How do you approach designing a class structure for a complex simulation?

Medium · Technical Fundamentals & Domain Knowledg

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?

Explain your understanding of AI behavior modeling.

Medium · Technical Fundamentals & Domain Knowledg

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 would you handle a memory leak or performance bottleneck in a simulation environment?

Medium · Technical Fundamentals & Domain Knowledg

Approach

Establish what changed and when, before forming any theory.
Pick a bisection that eliminates candidates whichever way it turns out.
Check the instrumentation before believing the symptom.
Separate the trigger from the cause; the deploy is rarely the bug.

Follow-up

What would you look at first, and what would it rule out?
How would you tell a cause from a coincidence here?

Suggested practice checklist

Built from the rounds and topics Discovery Machine candidates report.

Day 1 — Map the Discovery Machine loop

Write out the reported sequence: Online Application, Phone Screen, Technical Deep Dives, Whiteboard Challenges, Final Presentation.
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 5 reported rounds, with the weakest marked.

Day 2 — Work Artificial Intelligence (AI) Engineering

Spend the session on Artificial Intelligence (AI) Engineering, which Discovery Machine candidates report being tested on.
Write one worked example in Artificial Intelligence (AI) Engineering and time yourself on it.

Deliverable: One timed worked example in Artificial Intelligence (AI) Engineering.

Day 3 — Work Natural Language Processing (NLP)

Spend the session on Natural Language Processing (NLP), which Discovery Machine candidates report being tested on.
Write one worked example in Natural Language Processing (NLP) and time yourself on it.

Deliverable: One timed worked example in Natural Language Processing (NLP).

Day 4 — Work Speech Recognition

Spend the session on Speech Recognition, which Discovery Machine candidates report being tested on.
Write one worked example in Speech Recognition and time yourself on it.

Deliverable: One timed worked example in Speech Recognition.

Day 5 — Answer out loud: Technical Fundamentals & Domain Knowledge

Answer aloud, timed: Can you walk me through your experience with Java, C#, or Python?
Answer aloud, timed: How do you approach designing a class structure for a complex simulation?

Deliverable: Spoken answers to 2 reported Technical Fundamentals & Domain Knowledge question(s), under time.

Day 6 — Answer out loud: Behavioral & Cultural Fit

Answer aloud, timed: Tell me about a time you had to explain a complex technical solution to a non-technical stakeholder.
Answer aloud, timed: How do you handle feedback during code or model reviews?

Deliverable: Spoken answers to 2 reported Behavioral & Cultural Fit question(s), under time.

Day 7 — Dry run for Discovery Machine

Run one full mock under time, then write down the two questions you most want to ask your interviewers.

Deliverable: A completed timed mock and two questions to ask.

Behavioural preparation

Behavioural rounds judge the decision you made and what it cost.

Can you walk me through your experience with Java, C#, or Python?

Medium · Technical Fundamentals & Domain Knowledg

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?

What is your experience with relational databases like SQL Server or MySQL?

Medium · Technical Fundamentals & Domain Knowledg

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?