Full ownership of building predictive high-frequency U.S trading equities models, that generate millions of
dollars in annual profits.
Duties include developing new strategies, using machine learning algorithms and
statistical analysis, feature engineering, backtesting, fine tuning parameters and deploying on the live
Simplified model and addressed overfitting issues.
Efforts included removing redundant inputs, inputs with
weak correlation to the target, ranking and dropping inputs based on variance reduction.
complexity improved both profitability and trading system's timing metrics.
Improved model accuracy with ideas like ensemble models, cross validation, time-weighting samples, input
transformations, changing the error function, introducing synthetic samples.
Designed a tool to quantify 'missed opportunity' by analyzing market activity around our unfilled orders.
Output included potential profit outcome if we had filled and the speed up required to reach quotes on time.
This helped set both short and long term strategic goals for the team.
Wrote a 'Quote Attribution' framework that summarized the characteristics of the order book quotes our
orders interacted with.
Analyzing profitability along different dimensions led to insights for new features
and trading strategies.
Used Euclidean-distance based clustering to aggregate historical profit data from 'similar' stocks' and built
a regression framework to choose the 'best' model for a stock, from a group of candidates.
Performed detailed evaluation of a business proposition for distributed trading, analyzing the effect of
trading 'similar' models closer to each exchange.
Reported results to company CEO and team heads.
Wrote tools to test conformance between real-time trading and simulation results.
Keeping the conformance
gap small was crucial to having confidence in our simulation tests.
Tracking down conformance
discrepancies also helped uncover problems in the real-time trading environment.
First team member to switch from legacy C based trading platform to a Scala platform.
shared research tools to Scala, which eased and aided the transition for other team members.
Published key findings via email, Wiki and presentations at company wide meetings.
01/2008 - 01/2012
Quantitative Researcher, Software Developer
Ported the legacy execution software to a new framework in C , that allowed for a richer and more
flexible trading environment.
This helped make the system more adaptive and reduced operational risks
Implemented RegNMS required Intermarket Sweep Order logic in our trading systems that led to better and
This feature increased average daily profitability by at least 5%.
Led several end to end system integration and testing efforts.
Worked with the Compliance and
Development teams to design real-time monitoring tools to flag trading anamolies.
Assisted the Operations team in timely troubleshooting of real-time trading issues.
Designed and implemented the Triparty Repurchase Financing Allocation Server, a real-time
multi-threaded server, connecting the firm's trading system to Bank of New York's allocation system.
server handled allocation messages for the firm's multi-billion dollar portfolio.
Implemented an Execution Simulation Server to proxy for external counterparty servers, eliminating the
need for counterparty systems to be available when internal teams wanted to test.
Credit Technology | Financial Technology Associate 2004 - 2005
Implemented Boost Python Wrappers around C trading models, porting them to the Python environment.
This facilitated rapid testing of 'novel' ideas with Python's extensive statistical and graphical packages.
Education and Training
University of IllinoisChampaignM.S: Computer ScienceComputer Science
Birla Institute of Technology & Science (BITS)Pilani, IndiaM.Sc.(Tech): Information SystemsInformation Systems
Seeking senior data scientist/technical lead role in a dynamic, growth-oriented, collaborative environment.
Successful big data scientist and software programmer with broad skill sets and experience at fast-paced
Led the development, integration, testing and on-going maintenance of latency-sensitive, fault tolerant
Self-directed leader with proven ability to prioritize and deliver on multiple projects simultaneously.
Hands-on, goal-oriented problem-solver, with attention to detail and capable of deep diving into complex
Experience conducting research on open-ended and ill-defined topics.
Actively seek out new challenges and adapt easily to evolving technologies and concepts.
Excellent communication skills, adept at interacting with cross-functional teams. Held high-visibility roles
reporting directly to senior management. Mentored junior team members.