Internships represent a short period of your career. At Tower, we make sure this time is more than just a line on your resume. Make yourself stand out with an internship like no other.
It’s the most common question we get about our internship program – and a tricky one, because no two days are the same. You’ll be assigned to tasks and projects that align with your skillset and your team’s specific goals. Here’s a sense of what you can expect.
We’re always on the move – attending career fairs, hosting info sessions, and meeting future interns like you. Enter your email for updates on when Tower is visiting your campus next.
Whether you’re just beginning to explore quantitative finance careers or doing some last-minute interview prep, here are some key concepts all Tower interns should know.
Quantitative trading relies on mathematical models, algorithms, and data analysis to make trading decisions – rather than intuition or human discretion. Unlike discretionary trading, quant strategies are tested, automated, and executed programmatically.
A market order executes immediately at the best available price. A limit order sets a specific price at which you’re willing to buy or sell – it may not execute right away but gives you more control over price.
A quant researcher focuses on building trading models – identifying patterns, testing hypotheses, and generating signals. A quant developer builds the systems and infrastructure that bring those models to life – from data pipelines to execution engines. Both roles are essential, and they often work closely together.
Not necessarily. Many quants come from math, physics, CS, or engineering backgrounds. What matters most is analytical thinking, technical skill, and a curiosity about markets.
C++ and Python are among the most common. C++ is used for speed-sensitive systems, while Python is often used for research, modeling, and data analysis. Tower is also strategically investing in Rust, and is using it for multiple production systems within our core trading infrastructure.
Market microstructure refers to how trades actually happen – order books, matching engines, fees, latencies. Understanding it is key to building strategies that work in the real world.
A measure of performance over and above the market return. Generating alpha is the ultimate goal of most quant trading teams.
A measure of how much something moves relative to the market. A stock with beta > 1 is more volatile than the market; < 1 is less volatile.
Testing a trading strategy or model on historical data to evaluate how it would have performed before risking real capital.
A large pool of computers working together to handle heavy workloads. In quant trading, compute grids are used for backtesting, simulations, and data analysis that require massive parallel processing.
The change in value of a position for a change in price of the underlying instrument.
Strategies that trade spot FX, cash equities, bonds, or futures are generally referred to as delta one strategies, as a change in price of $1 in the instrument leads to a $1 change in profit and loss. This can be contrasted with options strategies, where the relationship is more complicated.
The method by which a trade is carried out – good execution strategies aim to minimize market impact, slippage, and latency while maximizing fill quality.
A metric that can be used as an input to a predictive signal. Features can be built from raw market or company data – for example, yesterday’s stock return, the difference between bid and ask prices, or a moving average. Good features help a model spot patterns that may predict future returns, risks, or trading opportunities.
When your order actually executes in the market. Traders often track fill rates (for example, by calculating the fill quantity divided by the order quantity).
Long-distance wireless communication using shortwave radio bands. In trading, HF links can connect markets across continents (like New York–London) with lower latency than undersea cables, though they are more weather-sensitive.
An algorithmic trading approach that relies on extremely fast execution speeds to capitalize on short-lived market opportunities.
The time delay between an instruction and its execution – in trading, lower latency often means better fill prices and more competitive performance.
Providing liquidity by continuously quoting both buy and sell prices for financial instruments – market makers help keep markets running smoothly.
A way to send trading data between cities using microwave radio links. Microwave is faster than fiber optic cables because signals travel more directly through the air, making it popular for low-latency trading between Chicago and New York.
A strategy that sits between high-frequency and long-term trading, focused on holding positions for hours to days using both speed and predictive models.
Turning a trading signal into real money. A signal by itself may just indicate “buy” or “sell” or forecast a future price. Monetization is about designing actual trades (size, timing, execution method) that make the signal profitable in practice.
Servers and infrastructure physically located in the firm’s own data centers. On-prem systems give firms more control, lower latency, and lower unit cost but require higher maintenance and up-front cost commitment than cloud-based systems.
The live system that actually runs trades in the market. Production is the final, fully tested environment where models and code interact with real money and real market data. Reliability is critical.
Trading a firm’s own capital (rather than client funds) to generate profit.
The difference between the expected price of a trade and the price at which it actually executes.
The difference between two prices.
A strategy that uses statistical models to exploit pricing inefficiencies between related securities – often implemented across hundreds or thousands of trades. Common statistical arbitrage strategies include mean reversion (profiting from temporary price dislocations that are expected to revert to the mean) and trend following (capitalizing on the continuation of price momentum).
A model- or data-driven trigger that suggests a buying or selling opportunity – signals are the foundation of most quantitative strategies.
A testing environment that mimics the real world but is separate from the production environment. UAT is where quants and developers check that systems behave as expected before moving them to production. Unlike prod, no real money is at risk.