What to Expect from a Tower Research Capital Interview
Tower Research Capital - typically just called Tower - is one of the largest and most established high-frequency trading firms in the world. Founded in 1998 by Mark Gorton, Tower trades equities, futures, options and FX across virtually every major exchange globally, with offices in New York, Chicago, London, Bangalore, Singapore and a handful of others. The firm runs a partnership-of-trading-teams model: each team operates relatively independently, with its own strategies and infrastructure under a shared umbrella. The interview process reflects this - you are interviewing both with the firm and with a specific team.
This guide covers Tower's process for software engineer, quantitative researcher and trader roles, the question types that come up, and a six-week preparation plan that maps onto the Quantt coding tests.
Tower Research Capital at a Glance
- Founded: 1998 by Mark Gorton
- Headquarters: New York, New York
- Size: ~1,300 employees globally
- Offices: New York, Chicago, London, Bangalore, Singapore, Geneva, Gurgaon, Charlotte
- What they trade: Equities, futures, options, FX (HFT and proprietary strategies)
- Roles they hire: Software Engineer, Quantitative Researcher, Trader, FPGA Engineer, Strategy Engineer
- Application route: careers page or via the Tower Research firm page on Quantt
The Interview Process
Stage 1: Online Assessment
For software engineer candidates, two or three algorithmic problems in 90 to 120 minutes. For quantitative researcher candidates, a probability and statistics test of similar length, often with a small data analysis component.
Stage 2: Phone Screens
One or two 60-minute calls. Live coding for engineers; probability and research methodology for researchers.
Stage 3: Onsite Super Day
Five to six interviews in a single day. Tower's distinctive feature: you usually meet representatives from two or three different trading teams, each with slightly different priorities. Decisions are partially team-dependent.
Stage 4: Team Match
After offers, candidates spend several days speaking with the teams who are interested. Compensation varies meaningfully across teams (some teams have higher base, others have higher bonus upside) so this stage is genuinely consequential.
How Interviews Differ by Role
Software Engineer
A FAANG-style technical interview with heavy emphasis on low-latency design. C++ is heavily favoured for hot-path systems; Python for platform work. Expect questions on lock-free data structures, the C++ memory model, kernel bypass networking, and order book design.
Typical split: 50% coding, 30% systems design, 10% C++ specifics, 10% behavioural.
Quantitative Researcher
A blend of statistics, signal design and Python. Tower's research culture is more team-specific than at Two Sigma or DE Shaw - the questions you face will reflect the specific strategies of the team interviewing you.
Trader
Tower's trading is heavily automated, so the trader role is closer to a senior researcher with operational responsibility. Expect questions on probability, market microstructure, and the practical realities of running automated strategies.
FPGA Engineer
A specialist track. Expect Verilog/VHDL, FPGA architecture, and questions about the trading-systems integration of FPGA designs.
Real Question Types
Coding
Question 1: Top-K updates Implement a service that maintains the top-K largest values among a stream where values can be updated (not just inserted). Aim for O(log N) per update.
Approach: A sorted data structure (e.g., balanced BST or Boost's multi_index) keyed by value, plus a hash map from key to current value. On update, find the old entry by hash map lookup, remove it, insert the new value. Maintain a pointer to the K-th largest for fast top-K reads.
Question 2: Branch prediction
You have a loop that iterates 1 million times and inside there is an if (rare_condition). The condition is true less than 1% of the time. What can you do to improve performance?
Approach: Three options. (1) Hint the compiler with __builtin_expect or C++20 [[unlikely]]. (2) Restructure to avoid the branch entirely (sometimes possible with arithmetic). (3) Profile-guided optimisation. Mention that on modern CPUs the compiler usually predicts well already, so measure first.
Systems Design
Question 3: Timestamping Design a system that timestamps incoming market data with nanosecond precision and millisecond-stable accuracy across machines. What are the trade-offs?
Approach: Hardware timestamping at the NIC (e.g., Solarflare PTP). Use IEEE 1588 PTP for clock sync across hosts (sub-microsecond accuracy). For the highest accuracy, GPS-disciplined clocks. Trade-offs: PTP requires switch support; GPS requires antenna and cabling; both add cost. Software timestamps via TSC are fast but suffer from CPU frequency scaling and core migration.
Probability
Question 4: Random walk hitting time A random walk starts at 0. At each step, it moves +1 with probability p = 0.6 or -1 with probability 0.4. What is the expected number of steps to first hit +10?
Approach: For an asymmetric random walk with positive drift, the expected hitting time of N is N / (p - q) = 10 / 0.2 = 50.
Behavioural
Question 5: Disagreement Tell me about a time you disagreed with a senior colleague about a system design decision. How did you handle it?
Approach: Tower values direct, evidence-based disagreement. Describe specifically: the design choice, why you disagreed, what evidence you brought, how it was resolved, what the outcome was. Avoid stories where you were "obviously right" - they want to see you grappling with a real trade-off.
How to Prepare - A Six-Week Plan
Weeks 1-2: Foundations. 100 LeetCode hard problems with a strict time budget. For research candidates, work through A Practical Guide to Quantitative Finance Interviews (Xinfeng Zhou) plus the first half of The Elements of Statistical Learning.
Weeks 3-4: Tower-specific depth. For engineers, Effective Modern C++ (Meyers), C++ Concurrency in Action (Williams), and a primer on Solarflare/onload and DPDK. For researchers, read about HFT signal design - Cartea, Jaimungal and Penalva's Algorithmic and High-Frequency Trading is the standard reference.
Weeks 5-6: Mock interviews. Three full mock onsites under realistic conditions.
For broader context, see our hardware acceleration for quant, networking fundamentals for developers, and latency arbitrage explained guides.
What Tower Looks For Beyond Technical Skill
Tower's team-of-teams culture means cultural fit is genuinely team-specific. That said, three traits matter across the firm.
Comfort with the team-match dynamic. Tower's structure means the firm cannot guarantee a specific role; candidates who can speak compellingly to multiple teams about why they would fit do better than candidates who pre-commit to one team.
Engineering rigour without academic pretension. Tower's culture is described as practical and engineering-driven - the firm is less academic than Jump or HRT. Candidates who can ship working systems quickly and iterate based on real performance data outperform candidates who emphasise theoretical elegance.
Long-term mindset. Tower has unusually long tenure for an HFT firm - many traders and engineers have been there a decade or more. The firm screens for candidates who are genuinely interested in the problem space rather than just the compensation.
For broader context, see our quantitative analyst career guide.
Compensation & recruiting notes
Pay ranges in this guide are illustrative estimates from publicly discussed bands and anecdotal reports - not official figures from the employer. Packages vary widely by desk, office, performance, and year. Hiring processes change; nothing here guarantees an interview, assessment format, or offer.
Frequently Asked Questions
How hard is it to get a Tower Research interview?
Competitive, broadly comparable to Jump and HRT. Tower hires roughly 80 to 120 graduates globally each year. Campus recruiting is concentrated at top US engineering programmes plus IIT for the Bangalore office.
Does Tower hire from non-target universities?
Yes, more readily than some peers because of the team-of-teams structure - some teams are more open to non-traditional backgrounds than others.
What programming languages should I know for a Tower interview?
C++ is heavily favoured for engineering roles. Python for research. Some teams use Rust or Go. The bar on whichever language you choose is high.
How does Tower's compensation compare to other HFT firms?
Tower sits in the upper tier. Graduate engineers typically receive $200,000 to $350,000 in their first year (base plus signing plus first-year bonus); the range is wider than at peers because compensation varies meaningfully across teams. Senior engineers and researchers earn into the seven figures. See our quantitative analyst salary guide.
How many interview rounds does Tower have?
OA, one or two phone screens, and a final-round Super Day with five to six interviews. After the offer, a team-match process with two or three groups.
What is the difference between Tower and Jump?
Both are large, established HFT firms with global footprints. Jump is more centralised and more proprietary about its strategies; Tower is structured as a partnership of more independent teams. Cultures differ accordingly - Jump is described as more uniform; Tower as more varied across teams.
Can I reapply if rejected?
Yes, after 12 months. Tower takes repeat candidates seriously, particularly those who can show measurable improvement.
Practise the questions Tower Research Capital Interview: Process, Questions and Prep 2026 actually asks
Reading about the interview is one thing - sitting one is another. Quantt's interactive coding tests are modelled on the same problem types that show up in firms like Jane Street, Citadel, Hudson River and Optiver. Run real Python in the browser, get instant feedback, and benchmark yourself against the bar.
Free to start - no credit card required