cswork / libraryOpen practice workspace

Interview guides · Bmll Technologies

Bmll Technologies Software Engineer Interview Guide

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

The 60-second brief

At Bmll Technologies, a Software Engineer plays a pivotal role in democratizing access to complex financial data. The company is a pioneer in historical financial data science, specializing in Level 3 order book data and analytics. As a engineer here, you are not simply building standard web applications; you are designing, optimizing, and scaling the backend systems and APIs that process billions of financial messages daily. Your work directly impacts global investment banks, quantitative hedge funds, and asset managers who rely on Bmll Technologies to backtest trading strategies, conduct market microstructure research, and comply with regulatory requirements. You will work on high-performance computing pipelines, cloud-native architectures, and data delivery systems that must handle petabyte-scale datasets with absolute precision and minimal latency.

This guide is scoped to a Software Engineer candidate at Bmll Technologies.

Focus areas

  • Python
  • Dynamic Programming
  • System Design

The Bmll Technologies interview process

Bmll Technologies candidates report 4 rounds over 3-5 weeks. The stages below are what candidates describe, not a published process.

Initial Conversation

A two-way discussion with the hiring manager about your background, engineering experience, and the role's scope.

What to demonstrate

  • A two-way discussion with the hiring manager about your background, engineering experience, and the role's scope
  • Depth in Python

How to prepare

  • Answer aloud and timed: Given a set of strings, write an optimized algorithm to find the longest common subsequence using dynamic programming.
  • Answer aloud and timed: Implement a memory-efficient parser for a stream of financial transaction logs.

Live Coding Session

A rigorous technical evaluation where you demonstrate your coding capabilities in real-time.

What to demonstrate

  • A rigorous technical evaluation where you demonstrate your coding capabilities in real-time
  • Depth in Python

How to prepare

  • Answer aloud and timed: Write a function to detect cycles in a directed graph representing dependency chains in a data pipeline.
  • Answer aloud and timed: How would you optimize a search algorithm to run in sub-linear time over a sorted but continuously updating dataset?

System Design Discussion

An in-depth discussion focused on designing systems and understanding architectural principles.

What to demonstrate

  • An in-depth discussion focused on designing systems and understanding architectural principles
  • Depth in Python

How to prepare

  • Answer aloud and timed: Explain how the Global Interpreter Lock (GIL) affects multi-threading in Python, and how you would bypass it for CPU-bound tasks.
  • Answer aloud and timed: Describe the difference between deep and shallow copying in Python, and how memory allocation is managed under the hood for large lists.

Deep-Dive Technical Interview

A thorough interview concentrating on language internals and specific engineering concepts.

What to demonstrate

  • A thorough interview concentrating on language internals and specific engineering concepts
  • Depth in Python

How to prepare

  • Answer aloud and timed: How do Python decorators work? Write a custom decorator that profiles memory usage and execution time of a function.
  • Answer aloud and timed: Explain the difference between generators and iterators, and demonstrate how generators optimize memory when processing large data files.

Where candidates lose points

Master Python's "Why

Do not just know how to write Python code; understand why a specific approach is faster or more memory-efficient. Be prepared to explain the inner workings of generators, memory allocation, and the GIL.

Focus on Trade-offs in System Design

There is no single "correct" architecture. When designing a system, always present multiple options and explain why you chose one over another based on throughput, latency, cost, and complexity.

Drive the Coding Collaboration

If you get stuck during a coding round, do not freeze up. Talk through your thought process, state what you are trying to achieve, and actively invite the interviewers to collaborate with you.

Highlight Data-Scale Experience

Throughout your interviews, emphasize any past experience you have working with large datasets, pipeline optimization, or high-performance computing, as this directly aligns with the company's core business challenges.

Questions & practice

Given a set of strings, write an optimized algorithm to find the longest common subsequence using dynamic prog

Medium · Coding and Algorithmic Problem Solving

Given a set of strings, write an optimized algorithm to find the longest common subsequence using dynamic programming.

Approach

  1. Restate the input: its shape, its size, and what is guaranteed about it.
  2. Name the brute-force solution and its complexity before improving on it.
  3. Choose the data structure from the access pattern, not from familiarity.
  4. State the target complexity and say which constraint rules the naive version out.

Follow-up

  • How does this change if the input no longer fits in memory?
  • What is the worst case, and how likely is it on real data?

Implement a memory-efficient parser for a stream of financial transaction logs.

Medium · Coding and Algorithmic Problem Solving

Approach

  1. Say what the runtime actually does before reasoning about the code.
  2. Name what is shared across threads and what owns each piece of state.
  3. Identify the window where an invariant is briefly untrue.
  4. Distinguish a value from a reference to it, and say which one you handed out.

Follow-up

  • What happens if two callers reach this at the same time?
  • Where could this allocate more than you expect?

Write a function to detect cycles in a directed graph representing dependency chains in a data pipeline.

Medium · Coding and Algorithmic Problem Solving

Approach

  1. Restate the input: its shape, its size, and what is guaranteed about it.
  2. Name the brute-force solution and its complexity before improving on it.
  3. Choose the data structure from the access pattern, not from familiarity.
  4. State the target complexity and say which constraint rules the naive version out.

Follow-up

  • How does this change if the input no longer fits in memory?
  • What is the worst case, and how likely is it on real data?

How would you optimize a search algorithm to run in sub-linear time over a sorted but continuously updating da

Medium · Coding and Algorithmic Problem Solving

How would you optimize a search algorithm to run in sub-linear time over a sorted but continuously updating dataset?

Approach

  1. Restate the input: its shape, its size, and what is guaranteed about it.
  2. Name the brute-force solution and its complexity before improving on it.
  3. Choose the data structure from the access pattern, not from familiarity.
  4. State the target complexity and say which constraint rules the naive version out.

Follow-up

  • How does this change if the input no longer fits in memory?
  • What is the worst case, and how likely is it on real data?

Explain how the Global Interpreter Lock (GIL) affects multi-threading in Python, and how you would bypass it f

Medium · Python Core and Internals

Explain how the Global Interpreter Lock (GIL) affects multi-threading in Python, and how you would bypass it for CPU-bound tasks.

Approach

  1. Say what the runtime actually does before reasoning about the code.
  2. Name what is shared across threads and what owns each piece of state.
  3. Identify the window where an invariant is briefly untrue.
  4. Distinguish a value from a reference to it, and say which one you handed out.

Follow-up

  • What happens if two callers reach this at the same time?
  • Where could this allocate more than you expect?

Describe the difference between deep and shallow copying in Python, and how memory allocation is managed under

Medium · Python Core and Internals

Describe the difference between deep and shallow copying in Python, and how memory allocation is managed under the hood for large lists.

Approach

  1. Say what the runtime actually does before reasoning about the code.
  2. Name what is shared across threads and what owns each piece of state.
  3. Identify the window where an invariant is briefly untrue.
  4. Distinguish a value from a reference to it, and say which one you handed out.

Follow-up

  • What happens if two callers reach this at the same time?
  • Where could this allocate more than you expect?

How do Python decorators work? Write a custom decorator that profiles memory usage and execution time of a fun

Medium · Python Core and Internals

How do Python decorators work? Write a custom decorator that profiles memory usage and execution time of a function.

Approach

  1. Say what the runtime actually does before reasoning about the code.
  2. Name what is shared across threads and what owns each piece of state.
  3. Identify the window where an invariant is briefly untrue.
  4. Distinguish a value from a reference to it, and say which one you handed out.

Follow-up

  • What happens if two callers reach this at the same time?
  • Where could this allocate more than you expect?

Explain the difference between generators and iterators, and demonstrate how generators optimize memory when p

Medium · Python Core and Internals

Explain the difference between generators and iterators, and demonstrate how generators optimize memory when processing large data files.

Approach

  1. Say what the runtime actually does before reasoning about the code.
  2. Name what is shared across threads and what owns each piece of state.
  3. Identify the window where an invariant is briefly untrue.
  4. Distinguish a value from a reference to it, and say which one you handed out.

Follow-up

  • What happens if two callers reach this at the same time?
  • Where could this allocate more than you expect?

Discuss the trade-offs between using SQL versus NoSQL databases for storing structured financial market events

Medium · System Design and Scalability

Discuss the trade-offs between using SQL versus NoSQL databases for storing structured financial market events.

Approach

  1. Name the grain you start from and join outward from it.
  2. Check whether any join is one-to-many before aggregating, or the sums inflate.
  3. Say which index the query would use, and what makes it unusable.
  4. 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?

Replace offset paging on the resource feed with keyset

Easy · keyset pagination · composite index · sort order

resource holds resource_id, tenant_id, owner_user_id, title, body_ref, version, status ('draft','active','archived','deleted'), created_at, updated_at, deleted_at, with an index on (tenant_id, status, updated_at DESC, resource_id DESC). The listing endpoint returns active resources for one tenant, newest update first, 50 per page, today with LIMIT 50 OFFSET n. Tenants reach page 400 and rows are created while they read. Write the keyset query, define what the cursor carries and how it is encoded, and say which part of the index each predicate uses. Assume PostgreSQL 16.

Approach

  1. Name the two failures separately. OFFSET 20000 makes the server produce and discard 20,000 rows, so page cost grows with depth rather than with page size. Independently, any write that changes how many rows sort above the offset moves the window between two fetches, and the direction decides which anomaly you get: an insert lands at the head of updated_at DESC and pushes already-returned rows down past the boundary, so they are returned a second time; a delete above the offset, or a row whose updated_at is bumped above the cursor, pulls rows up and one is never returned at all. Nothing in the response reveals either.
  2. Write the seek: WHERE tenant_id = $1 AND status = 'active' AND (updated_at, resource_id) < ($2, $3) ORDER BY updated_at DESC, resource_id DESC LIMIT 50. The row-value comparison is one index range rather than a disjunction, and both columns are NOT NULL, which is what makes that comparison well defined.
  3. Map each predicate onto the index: tenant_id and status are equality on the leading columns, (updated_at, resource_id) is the range, and the ORDER BY matches the index order so no Sort node appears and the scan stops after 50 rows. The DESC in the definition only matters for mixed directions — a plain ascending btree on the same columns is read backwards for this query.
  4. Put both sort columns in the cursor and nothing the client can tamper with into another tenant: base64 of (updated_at, resource_id), validated server-side, with tenant_id taken from the principal.

Follow-up

  • The client asks for 'jump to page 400'. What do you offer instead, and what does the honest version cost?
  • Sort order becomes user-selectable across four columns. How many indexes is that, and which would you refuse to add?

What are dunder (double underscore) methods, and how would you use them to customize class instantiation or op

Medium · Python Core and Internals

What are dunder (double underscore) methods, and how would you use them to customize class instantiation or operator overloading?

Approach

  1. Clarify what is being asked and what a complete answer contains.
  2. State your assumptions explicitly before working the problem.
  3. Say what you would check first and why it is the highest-information step.
  4. 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 design a distributed, fault-tolerant system to ingest and process gigabytes of financial market

Medium · System Design and Scalability

How would you design a distributed, fault-tolerant system to ingest and process gigabytes of financial market data per second?

Approach

  1. Fix the scope first: who calls this, how often, and what they do when it fails.
  2. Name the read and write paths separately; they rarely have the same bottleneck.
  3. Choose a partition key and say what query it makes expensive.
  4. 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?

Describe your approach to scaling a RESTful API that serves historical data queries to thousands of concurrent

Medium · System Design and Scalability

Describe your approach to scaling a RESTful API that serves historical data queries to thousands of concurrent institutional clients.

Approach

  1. Say who the caller is and what they do when the call fails halfway.
  2. Define the identity of a request so a retry cannot double-apply it.
  3. Separate accepted, pending, failed and confirmed; they are different facts.
  4. 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?

How would you design a caching layer to minimize database load for frequently accessed historical order book s

Medium · System Design and Scalability

How would you design a caching layer to minimize database load for frequently accessed historical order book states?

Approach

  1. Fix the scope first: who calls this, how often, and what they do when it fails.
  2. Name the read and write paths separately; they rarely have the same bottleneck.
  3. Choose a partition key and say what query it makes expensive.
  4. 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?

Edge instances grow 400 MB per hour until the nightly restart

Hard · memory leak · heap analysis · metric cardinality

Edge API instances start at 700 MB resident and grow about 400 MB/hour; a nightly rolling restart has hidden it for weeks. Growth continues unchanged when request rate halves overnight, p99 degrades in the last hours before an instance is recycled, and heap used immediately after a forced full GC rises monotonically. The service holds no product state. Name the discriminating measurement that separates the plausible causes, give the most likely cause, and give the fix and how you would verify it.

Approach

  1. Separate resident memory from live heap first, because they fail differently. Resident size can grow from fragmentation, native buffers or thread stacks while the heap is flat; heap used after a full GC rising monotonically is the measurement that says objects are reachable and not being released. You already have it, so this is retention, not fragmentation, and that closes off half the candidate list.
  2. Use the rate's independence from traffic as the discriminator. Growth that continues at half the request rate rules out per-request objects that are merely slow to collect and points at a structure that grows with distinct values observed rather than with call volume. Write the candidates that have that property: a metrics registry keyed on a high-cardinality label, an unevicted cache, an interner, a per-key lock map.
  3. Take two heap snapshots an hour apart and diff by retained size, reading the dominator tree, not by allocation count or instance count. Expect one root holding a map with millions of entries, then follow the reference chain to the code that inserts and never removes. Allocation profilers point at churn, which is the wrong signal here.
  4. The candidate that fits this service is an observability label carrying an identifier, such as a request path recorded before templating so that /v1/resources/48213 becomes its own metric series. That grows with distinct ids seen, is independent of rate, and explains the late p99 degradation, since GC cost rises with the size of the live set.

Follow-up

  • Post-GC heap is now flat but resident size still creeps. What are you looking at, and does it matter?
  • How would you have detected this before an OOM, given the nightly restart masked the trend?

Suggested practice checklist

Built from the rounds and topics Bmll Technologies candidates report.

Day 1 — Map the Bmll Technologies loop

  • Write out the reported sequence: Initial Conversation, Live Coding Session, System Design Discussion, Deep-Dive Technical 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.

Day 2 — Work Python

  • Spend the session on Python, which Bmll Technologies candidates report being tested on.
  • Write one worked example in Python and time yourself on it.

Deliverable: One timed worked example in Python.

Day 3 — Work Dynamic Programming

  • Spend the session on Dynamic Programming, which Bmll Technologies candidates report being tested on.
  • Write one worked example in Dynamic Programming and time yourself on it.

Deliverable: One timed worked example in Dynamic Programming.

Day 4 — Work System Design

  • Spend the session on System Design, which Bmll Technologies candidates report being tested on.
  • Write one worked example in System Design and time yourself on it.

Deliverable: One timed worked example in System Design.

Day 5 — Answer out loud: Coding and Algorithmic Problem Solving

  • Answer aloud, timed: Given a set of strings, write an optimized algorithm to find the longest common subsequence using dynamic programming.
  • Answer aloud, timed: Implement a memory-efficient parser for a stream of financial transaction logs.

Deliverable: Spoken answers to 2 reported Coding and Algorithmic Problem Solving question(s), under time.

Day 6 — Answer out loud: Python Core and Internals

  • Answer aloud, timed: Explain how the Global Interpreter Lock (GIL) affects multi-threading in Python, and how you would bypass it for CPU-bound tasks.
  • Answer aloud, timed: Describe the difference between deep and shallow copying in Python, and how memory allocation is managed under the hood for large lists.

Deliverable: Spoken answers to 2 reported Python Core and Internals question(s), under time.

Day 7 — Answer out loud: System Design and Scalability

  • Answer aloud, timed: How would you design a distributed, fault-tolerant system to ingest and process gigabytes of financial market data per second?
  • Answer aloud, timed: Describe your approach to scaling a RESTful API that serves historical data queries to thousands of concurrent institutional clients.

Deliverable: Spoken answers to 2 reported System Design and Scalability question(s), under time.

Behavioural preparation

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

Walk me through the most technically challenging backend system you have designed and deployed to production.

Medium · Experience and Behavioral

Approach

  1. Pick a story where you made the decision, not one where you watched it.
  2. State the situation in two sentences and spend the rest on the reasoning.
  3. Give the blast radius: what could have broken, and what you measured.
  4. 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?

Explain a scenario where you had a technical disagreement with a teammate or lead. How did you resolve it and

Medium · Experience and Behavioral

Explain a scenario where you had a technical disagreement with a teammate or lead. How did you resolve it and arrive at a solution?

Approach

  1. Pick a story where you made the decision, not one where you watched it.
  2. State the situation in two sentences and spend the rest on the reasoning.
  3. Give the blast radius: what could have broken, and what you measured.
  4. 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?

Why do you want to work at Bmll Technologies, and how does your background in backend engineering align with o

Medium · Experience and Behavioral

Why do you want to work at Bmll Technologies, and how does your background in backend engineering align with our data-intensive product?

Approach

  1. Pick a story where you made the decision, not one where you watched it.
  2. State the situation in two sentences and spend the rest on the reasoning.
  3. Give the blast radius: what could have broken, and what you measured.
  4. 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 time when you had to optimize a slow-performing system. How did you identify the bottleneck, and wh

Medium · Experience and Behavioral

Describe a time when you had to optimize a slow-performing system. How did you identify the bottleneck, and what was the outcome?

Approach

  1. Pick a story where you made the decision, not one where you watched it.
  2. State the situation in two sentences and spend the rest on the reasoning.
  3. Give the blast radius: what could have broken, and what you measured.
  4. 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?