Quantitative Developer - Equities Research Engineering

NJF Global Holdings Ltd; New York (NY); On-site; USD 150,000 - 230,000; Full time

Job summary

NJF Global Holdings Ltd is seeking an engineer to join the New York research engineering group, building the intraday equities stack. You will work between strategy research and core development, creating backtesting simulators, data pipelines and reconciliation tooling that test, tune and check researchers' models.

You will focus on fast, clean C++ with attention to memory layout and cache behavior, delivering tools that keep training and live inference in step and surface PnL, volume and price

Qualifications

• Advanced C++ development with focus on low-latency systems.
• Experience with event-driven backtesting or tick-level simulation is a plus.
• Strong understanding of market microstructure signals and order book concepts.

Responsibilities

• Build and maintain the intraday equities research stack used by researchers.
• Develop a tick-by-tick backtesting simulator and data replay systems.
• Ensure production reconciliation between live trading state and captured market data.

Skills

Concurrency concepts

Cache optimization

Memory locality

Feature pipelines design

Statistical testing basics

Backtesting and simulation

P&L interpretation knowledge

Data analysis for tick data

Tools

C++

Tick data storage & replay

OLS regression

XGBoost

LightGBM

PCA

Data pipeline tooling

Job description

The firm trades equities, options and futures with its own capital on exchanges around the world, with offices across the US, Europe and Asia. Its core expertise is high-frequency trading: low-latency market making, arbitrage and market taking built on market microstructure signals. It is now building out mid-frequency strategies with holding periods from intraday to a few days, and the main focus of this build is intraday trading above the HFT horizon.

This role sits in the New York research engineering group that supports the firm's intraday equities build. The group sits between strategy research and core development: researchers own the signals and strategy logic, core development owns market data handling, order gateways and the live trading system, and this group builds the tools researchers use to test, tune and check their work, from simulation and data pipelines to optimization and reconciliation. The intraday build is new, so much of the group's work extends a stack designed for HFT to longer holding periods, and its output sets how quickly new strategies move from research to live trading.

You will build and maintain the research stack equities researchers use every day: an event-driven backtesting simulator that replays tick-by-tick market data, the storage and replay systems behind it at hundreds of gigabytes per day, and order book reconstruction from each exchange's add, cancel and trade messages. You will own production reconciliation, checking that a strategy's internal state in live trading matches calculations on captured market data and alerting on any gap. You will also build parameter optimization tooling (regressions for signal weights and hedge ratios, backtest sweeps), modular data cleaning and feature pipelines that keep training and live inference in step, and plotting tools for PnL, volume and price. You will sit with researchers, work across the full codebase, and see the PnL of the strategies your tools support.

Who they're looking for

We are seeking an engineer who writes fast, clean C++ and treats memory layout and cache behavior as part of the design, measuring each change rather than assuming. You should have close attention to detail, enough statistics to follow how researchers build and test models, and an interest in how equity markets work at the tick level. You will stand out if you have built a tick-level simulator, a market data capture or replay system, or tooling that took research models into production, or if you know equities microstructure signals such as book pressure, trade impulse and ETF constituent pricing.

Skills and expertise

• Concurrency (lock-free queues, multithreading, shared memory)
• Cache and memory-locality optimization
• Order book design and reconstruction
• Tick data storage and replay
• Streaming statistics in constant memory
• Backtesting and simulation (incl. Sharpe ratio evaluation)
• Regression and inference (OLS, hypothesis testing, z- and t-tests)
• Statistical learning (regularization, PCA, xgboost, LightGBM)
• Data cleaning and feature pipelines