Niranjan Smitha
Software Engineer · Algorithms · Finance · IIT Bombay
About Me
I am a Mechanical Engineering undergraduate at IIT Bombay with a strong interest in algorithmic trading, quantitative finance, and applied machine learning. My work focuses on building and evaluating systematic trading strategies, implementing financial models, and developing efficient algorithms for real-time data analysis. I enjoy working on problems that require mathematical rigor, careful modeling, and performance evaluation, and I am motivated by roles that emphasize analytical depth and disciplined problem-solving.
Skills
Projects
Trading Strategy Development (Forex – NY Session)
Problem: Discretionary trading decisions without systematic validation lead to inconsistent performance.
Approach: Designed and backtested a proprietary forex trading strategy tailored for the New York session, achieving a 71% win rate over a full month. Analyzed equity curves, P&L, drawdowns, and volatility behavior using structured trade logs and visual reports.
Options Pricing Models (Quantitative Finance)
Problem: Option values vary non-linearly with market parameters and require robust mathematical modeling.
Approach: Implemented Black-Scholes, Binomial, and Monte Carlo models in Python to price European options. Automated pricing across parameter ranges, generated CSV outputs, and documented theoretical foundations and sensitivity analysis.
Machine Learning on Real-Time Market Data (NIFTY-50)
Problem: Batch ML models struggle to adapt to continuously evolving financial time-series data.
Approach: Built an online ML system using stochastic gradient descent to predict NIFTY-50 price movements from streaming data. Engineered OHLC, volume, and temporal features; achieved test MSE ≈ 0.08 with efficient real-time performance.
Rubik’s Cube Solver using Korf’s IDA*
Problem: Brute-force search is infeasible for solving combinatorial state-space problems efficiently.
Approach: Modeled a virtual 3×3 Rubik’s Cube in C++ and implemented BFS, DFS, IDDFS, and Korf’s IDA* algorithm. Achieved solutions for 13-move scrambles in under 10 seconds using admissible heuristics.
Vehicle Routing Optimization using Genetic Algorithms
Problem: Finding optimal routes in VRP is NP-hard and unsuitable for exact methods at scale.
Approach: Applied genetic algorithms using the DEAP library to solve VRP. Designed a custom fitness function, evolved route populations, and visualized optimized solutions using Matplotlib.
Contact
Feel free to reach out for opportunities, collaborations, or discussions.