ALGORITHMIC & COMPUTATIONAL WORK

Technical Projects

Systematic portfolio optimization, stochastic calculus in derivative valuation, learned-dynamics HyperMPC at 100 Hz, physics-informed neural networks (PINNs), and multi-agent astronomical exploration.

QUANT // ALPHA

Effects of Interest Rate Hikes on Financial Markets & Vasicek Modeling

FINSEARCH, Finance Club | IIT Bombay

• Jun 2026 – Jul 2026

Empirical analysis of 20+ years of RBI monetary policy tightening cycles and simulation of bond price sensitivity and yield curves using the Vasicek short-rate model.

#Quantitative Finance #Fixed Income #Vasicek Model #Macroeconomics #Monetary Policy #Stochastic Rates

VIEW TECHNICAL DETAILS →

QUANT // ALPHA

Regime-Aware Portfolio Optimization

Summer of Quant, Quant Club | IIT Bombay

• Jun 2026

Built a regime-switching portfolio optimizer combining Gaussian HMMs, Random Forest meta-models and convex allocation across 503 equities.

#Quantitative Finance #Hidden Markov Models #Convex Optimization #CUSUM Sampling

VIEW TECHNICAL DETAILS →

ML // NEURAL

The Cosmic Detective: Multi-Agent AI for Sky Exploration

Krittika Summer Projects, Krittika Astronomy Club | IIT Bombay

• May 2026

Multi-agent AI system for autonomous astronomical exploration and hypothesis generation. Built Gemini-driven literature benchmark, Neo4j knowledge graph with 95% similarity de-duplication, and Swift FITS pipeline.

#Multi-Agent AI #LLM Orchestration #Neo4j Knowledge Graph #Astrophysics Data #Swift FITS Pipeline #Gemini API

VIEW TECHNICAL DETAILS →

ML // NEURAL

Inverse Elasticity Physics-Informed Neural Network (PINN)

ME218 Course Project | Guide: Prof. Sripriya Ramamoorthy, IIT Bombay

• Mar 2026 – Apr 2026

Direct Hybrid Collocation PINN estimating Young’s modulus fields across 3 Additive Manufacturing lattice specimens. Unified 8 DIC anchors with 10,000+ Sobol points enforcing 2D Cauchy Momentum PDEs.

#Machine Learning #Physics-Informed Neural Networks #Cauchy Momentum PDE #Inverse Problems #Sobol Collocation #Continuum Mechanics

VIEW TECHNICAL DETAILS →

ML // NEURAL

Single Image Super Resolution with Deep Residual Networks

ME228 Course Project | Guide: Prof. Neeraj Kumbhakarna, IIT Bombay

• Mar 2026 – Apr 2026

Enhanced Deep Super-Resolution (EDSR) with 32 residual blocks and PyTorch AMP. Achieved 34.84 dB PSNR and 0.939 SSIM on DIV2K, and 37.47 dB PSNR benchmarking SRCNN on Sen2Venus satellite imagery.

#Deep Learning #Computer Vision #Super Resolution #EDSR #PyTorch AMP #Satellite Imagery

VIEW TECHNICAL DETAILS →

AUTONOMY // MPC

Spatial Active Noise Control via Virtual Acoustic Sensing

ME791 Course Project | Guide: Prof. Sripriya Ramamoorthy, IIT Bombay

• Mar 2026 – May 2026

Virtual sensing framework for Spatial Active Noise Control using 4 remote microphones and an ObsTasNet deep observation-filter on 400 simulated acoustic scenes, achieving 12.6x faster pressure estimation.

#Acoustic Signal Processing #Spatial ANC #ObsTasNet #Deep Learning #Virtual Sensing #Wave Physics

VIEW TECHNICAL DETAILS →

AUTONOMY // MPC

Spatial Hyper-MPC with Neural Policy Distillation

ME444 Course Project | Guide: Prof. Seshu Pasumarthy, IIT Bombay

• Mar 2026 – Apr 2026

Hybrid Neural-MPC controller for 6-DOF active suspension using OSQP and Joint EKF. Distilled a 40-step MPC into a 4,736-parameter neural policy via 20k expert trajectories, reducing peak angular velocities up to 3x.

#Autonomous Controls #Model Predictive Control #Neural Policy Distillation #OSQP Solver #Joint EKF #6-DOF Vehicle Dynamics

VIEW TECHNICAL DETAILS →

Wildlife Hotspot Detector: Active Learning on Skewed Datasets

DS203 Course Project | Guide: Prof. Vinay Kulkarni, IIT Bombay

• Oct 2025 – Nov 2025

Engineered HOG/GLCM/LAB feature vectors, Random Forest dimension reduction (750+ features), and an Active Learning loop with SMOTE and XGBoost GridSearchCV, attaining 0.85 F1-Score on highly skewed data.

#Machine Learning #Active Learning #Feature Engineering #XGBoost #SMOTE #Computer Vision

VIEW TECHNICAL DETAILS →

Chicago Quant Alley: Crypto Trading Simulator & Optimizer

Seasons of Code, Web and Coding Club | IIT Bombay

• May 2025 – Jul 2025

Engineered a mid-frequency Bitcoin options/futures simulator with slippage modeling, async ccxt API execution, dynamic strike selection, and Multi-Armed Bandit hyperparameter optimization.

#Quantitative Finance #Crypto Derivatives #Options Trading #Multi-Armed Bandits #Async CCXT #Backtesting Engine

VIEW TECHNICAL DETAILS →

$Mathematics of Derivative Pricing & Risk Neutral Valuation preview$

Mathematics of Derivative Pricing & Risk Neutral Valuation

Summer of Science, Maths & Physics Club | IIT Bombay

• May 2025 – Jul 2025

Theoretical derivation and computational implementation of the Black-Scholes PDE, martingale pricing, Greeks sensitivity analysis, and Monte Carlo option pricing engines in Python.

#Quantitative Finance #Black-Scholes PDE #Stochastic Calculus #Greeks Sensitivity #Martingale Pricing #Monte Carlo Simulation

VIEW TECHNICAL DETAILS →

SSMphony: Linear-Complexity Audio Generation with Mamba & State Space Models

Independent Research

• Jan 2025

Explored linear-complexity State Space Models (SSM) and Mamba architectures for long-horizon audio synthesis, overcoming the quadratic attention bottlenecks of standard Transformer architectures.

#Deep Learning #State Space Models #Mamba #Audio Generation #Linear Attention #Sequence Modeling

VIEW TECHNICAL DETAILS →