Prathamesh Pilkhane

Software Engineer

Mumbai, Maharashtra, India0 mo experience

Key Highlights

Expert in low-latency trading systems.
Proficient in C++ and Python integration.
Strong background in computational geometry and NLP.

Stackforce AI infers this person is a Fintech specialist with expertise in low-latency systems and algorithmic trading.

Contact

Skills

Core Skills

Low Latency TradingC++High Frequency TradingCompilersComputational GeometryNatural Language Processing (nlp)

Other Skills

ANTLRC#Competitive ProgrammingDeep LearningFlaskJavaScriptKernel BypassingLow latency C++Machine LearningMultithreadingParallel ProgrammingParsingProfile Guided OptimisationProfilersProof of Concept Development

Experience

Daksh trading

Quant Software Developer

Oct 2024 – Present · 1 yr 5 mos · Hong Kong SAR · On-site

Designed and implemented a lightweight, zero-payload heartbeat framework to detect process failures and broadcast them efficiently across all running processes, improving system reliability and observability.
Leveraged kernel-bypassing techniques for low-latency networking on ExaNIC X25 cards using ExaSock drivers and the libexanic library. Extended this work to implement a preloading mechanism, enabling encrypted messages to be pre-placed on the NIC for rapid transmission.
Built a delay model to compute round-trip times from production data, integrating these values into simulations. This enhanced the accuracy of backtesting by better aligning simulated behavior with real-world latency conditions.
Contributed to the front-end development of an auto-hedging linear solver, adhering to constraints on option Greeks (e.g., delta, vega) and notional limits—significantly simplifying and accelerating end-of-day hedging operations.
Developed a proof of concept to integrate Python into a large-scale C++ codebase using pybind11, empowering researchers and traders to rapidly prototype and test ideas without developer intervention.
Mentored interns on advanced profiling techniques, including the use of profilers as well as on the usage of Profile-Guided Optimization (PGO) with GCC, pushing performance beyond conventional -O3 optimization levels.
Diagnosed and resolved key inefficiencies in simulation workflows, resulting in up to a 20% improvement in simulation time.

Real Time System DesignKernel BypassingLow Latency TradingLow latency C++Parallel ProgrammingSystems Design+3

Nk securities research

Software Developer Internship

May 2023 – Jul 2023 · 2 mos · Gurugram, Haryana, India · On-site

Responsible for developing faster & space efficient optimisations to programs involved in High Frequency Trading (HFT) strategies, a form of algorithmic trading involving very high speed decisions in stock market.
Developed a Proof of Concept of multithreaded program for faster simulations, used in testing out various strategies, handling synchronisation of threads in critical sections and avoiding deadlocks.
Tested various Profilers, for benchmarking programs, identifying part of the code responsible for slower execution.
Optimised the calculations of Implied Volatility using Black Scholes Model, removing redundant calculations while avoiding cache misses by using various profilers such as AMDμProf , heaptrack and easyprofiler

Proof of Concept DevelopmentProfilersHigh Frequency TradingMultithreading

Finiq consulting

Development Intern

Nov 2022 – Dec 2022 · 1 mo · Pune, Maharashtra, India · On-site

Tackled the problem of compiling a non type-casted language FinSPL to a type casted language, C#.
Enhanced a user friendly language, FinSPL to help users to write custom functions in a Python like appearance.
Developed the parser and compiler for the above mentioned language using Antlr in C#, converting FinSPL to C#.

CompilersParsingC#ANTLR

Technische universität braunschweig

Research Intern

May 2022 – Jul 2022 · 2 mos · Hannover-Braunschweig-Göttingen-Wolfsburg Region · On-site

Developed a successful and efficient strategy for The Homesteading Problem, and simulated the same algorithm.
Worked on a wide variety of problems including Voronoi games and their variations, Computational Geometry
problems such as packing and covering of geometrical shapes, and the Art Gallery Problem variants.
Developed upon the worst case optimal packing density of cubes in sphere, using the proof for cubes in cube.

Computational Geometry

Mizuho

Machine Learning Intern

Nov 2021 – Dec 2021 · 1 mo · Mumbai, Maharashtra, India · Remote

Developed a Proof of Concept for the task of extracting information from recorded phone calls of customers.
Tested out various modules and libraries, used Speech recognition library for transcribing the phone call along with
spaCy, an NLP library for extracting details from the above generated textual format of speech.
Implemented an Amount Checker code, which could check the correctness of records in database corresponding to
amount in words to that in figures. Developed it to be less prone to errors due to misspelt or incomplete words.

Machine LearningDeep LearningNatural Language Processing (NLP)Speech Recognition