Citadel is one of the world's premier hedge funds and market makers, operating across systematic trading, discretionary macro, and credit strategies. Citadel's India operations include a Mumbai presence for trading operations and quantitative research, and Bengaluru for technology development. Citadel and its market-making arm Citadel Securities together employ over 3,000 people globally, with the India team growing steadily as a centre for technology and quant talent.
Citadel India hires across two primary tracks. Software Engineering: Citadel's technology team builds the world's most sophisticated trading infrastructure. In India, SWEs work on real-time risk systems (monitoring portfolio risk across thousands of positions in microseconds), trading platform infrastructure (order routing, execution management, connectivity to global exchanges), and data engineering pipelines (processing terabytes of market data daily for research consumption). The dominant languages are C++ for performance-critical systems, Python for data analysis and research tooling, and Java for some operational systems. Linux systems expertise is required for most C++ roles: memory management, concurrency patterns (lock-free data structures, SIMD optimisation), and network programming. Quantitative Research: Citadel's quant researchers develop systematic trading signals using statistical and machine learning techniques. The India quant team focuses on Asian market strategies, alternative data analysis (satellite imagery, web scraping, credit card transaction data), and risk modelling. Python with PyTorch and scikit-learn is the primary ML stack.
Citadel's interview process is among the most demanding in the industry, typically running 4-5 rounds. Round 1 (Online Coding Screen): 90-120 minutes. Two to three coding problems in C++ or Python at LeetCode medium-to-hard level. Citadel's online assessment frequently includes problems requiring concurrent programming (thread safety, mutex, condition variable patterns) or memory management (implementing a memory pool, custom allocator). For quant researcher roles, the coding screen also includes Python statistical programming tasks. Round 2 (Technical Interview 1, Algorithms): 60-75 minutes. Problems at LeetCode medium-to-hard level with significant discussion of time and space complexity. Common themes: graph algorithms (shortest path variants, minimum spanning tree), dynamic programming (interval scheduling, optimisation problems), and string manipulation (parsing financial instrument symbols, matching order books). Round 3 (Technical Interview 2, Systems or Quant Math): For SWE roles: a systems round on low-latency C++ patterns, distributed systems (CAP theorem, Raft consensus basics), and Linux performance profiling. For quant roles: probability and statistics (conditional expectation, Bayesian inference, stochastic processes basics), market microstructure, and portfolio construction concepts. Round 4 (System Design): For senior roles. 'Design a real-time risk monitoring system that calculates Greeks (delta, gamma, vega) for a portfolio of 100,000 options positions in under 100 milliseconds'. High-performance systems thinking. Round 5 (Final/Culture Round): With a senior team member or Partner. Intellectual curiosity, long-term commitment, and culture fit are evaluated.
Citadel's SWE interviews distinguish themselves from product company interviews through the depth of low-latency C++ knowledge expected. Key concepts: Memory Management: stack vs heap allocation, RAII (Resource Acquisition Is Initialisation), smart pointers (std::unique_ptr vs std::shared_ptr trade-offs), and custom memory allocators (why and how to implement a slab allocator for low-latency contexts). Concurrency: std::thread, std::mutex, std::condition_variable, lock-free data structures (std::atomic, compare-and-swap CAS operations), memory ordering (std::memory_order_relaxed vs seq_cst), and thread pool patterns. Performance Optimisation: cache locality (struct-of-arrays vs array-of-structs for SIMD), branch prediction (branchless code patterns), and profiling tools (perf, gprof, valgrind). For quant research roles, Python performance tools matter more: NumPy vectorisation, Cython for hot paths, Numba JIT compilation for numerical computation. Common interview question: 'You have a sorted array of 1 billion integers. You need to find all pairs that sum to a target. How do you do this in under 1 second on a single machine with 32GB RAM?' This combines algorithmic thinking, memory constraint reasoning, and C++ implementation detail.
Citadel India compensation is among the highest in the Indian technology market, particularly for quant roles. 2026 benchmarks: Software Engineer (SDE, 0-3 years fresher/entry): Rs 40-70 LPA total comp. Mid-level SWE (3-6 years): Rs 70-130 LPA. Senior SWE (6+ years): Rs 120-200 LPA. Quantitative Researcher (entry, fresh PhD or quant master's): Rs 60-120 LPA base, with significant variable component. Senior Quant Researcher: Rs 150-400+ LPA in a strong year (the variable component at senior levels can be 2-5x base). Citadel's compensation is structured as base salary plus a discretionary annual bonus: in strong years, bonuses can significantly exceed base. The culture is high performance and high compensation: Citadel has strict performance review processes and manages out underperformers quickly relative to corporate tech companies.
Citadel India recruits from a narrow campus set: IIT Bombay, IIT Delhi, IIT Madras, IISc, and CMI for quantitative roles, and the same IITs plus BITS Pilani for software engineering. Off-campus applications are accepted but conversion rates are lower without a strong referral or a competitive programming profile on Codeforces (Candidate Master or higher). Experienced hires come predominantly from D.E. Shaw, Goldman Sachs Strats, and Tower Research India. Citadel's Gurugram office handles most India hiring; the typical timeline from first contact to offer is four to eight weeks. Candidates with gaps between rounds should use the time to sharpen C++ and probability prep rather than broadening scope.
