Back to Home

PoliStock

January 2025 – Present
FinTechSaaSAIReal-time SystemsNLP
  • Built a financial intelligence SaaS platform democratizing institutional-grade market insights for retail investors, backed by Gener8tor's gBETA program.
  • Developed Natural Language Orders (NLO) system allowing users to set buy/sell triggers using plain English with automatic execution across 27+ major brokerages.
  • Architected real-time event monitoring system analyzing 5,000+ stocks across 20+ volatility sources (SEC filings, Reddit, financial news) with sub-1-minute alert delivery.
  • Scaled to 200+ active users with $250k+ invested based on platform insights, achieving 4.8-star rating with bank-level encryption and accessible $9.99/month pricing.

Spectator AI - Sports Analytics Platform

August 2022 – Present
Computer VisionDeep LearningAzureMLOps
  • Secured USD 25000 in Azure credits and USD 2500 in OpenAI API credits; architected scalable ML infrastructure using Azure ML and Docker containers.
  • Implemented computer vision pipeline with YOLOv11 for player tracking with 95% mAP; built transformer-based models with BERT fine-tuning for sports content analysis.
  • Designed tactical analysis system using Grounding DINO and trajectory clustering algorithms to identify formation patterns and strategic movements.
  • Created production ML pipeline with MLflow for experiment tracking and FastAPI microservices for inference deployment.

AMP®-Parkinson's Disease Progression

2023
Random-ForestsNeural-NetworksStatistical-AnalysisResearch
  • Predicted the progression of Parkinson's disease using multiple machine learning models to achieve maximum accuracy.
  • Applied advanced statistical analysis and neural network architectures to identify key progression indicators.

IPL Cricket Index

2023
Data-AnalyticsExcelData-ScienceData-Visualisation
  • Developed a comprehensive cricket batting index metric that provides a single measure of a batsman's performance.
  • Analyzed multiple aspects of batting ability including run rate, consistency, and performance against varying bowling styles.

FIFA-PREDS

2022
Data-AnalyticsExcelData-ScienceData-Visualisation
  • Analyzed team performance data from FIFA World Cup matches to provide comprehensive insights.
  • Built interactive visualizations to track performance metrics and identify success patterns across tournaments.

Covid-19 Spread Analysis

2022
Data-VisualisationClusteringCorrelationsBayesian-Analysis
  • Created a program that highlights peak spread periods of Covid-19 in each country through visualization and clustering.
  • Implemented Bayesian polynomial regression to analyze correlations between different spread factors.

Zillow Housing Prices

2022
Linear-RegressionStatistical-AnalysisNeural-Networks
  • Used Linear Regression to provide accurate 'Zestimate' property valuations.
  • Performed comprehensive statistical analysis on large real estate datasets to identify pricing trends and indicators.

NASA Glacier Tracking System

November 2024 – Present
Climate ScienceGeospatialPythonCUDABig Data
  • Built an automated glacier tracking system using Xarray, NumPy, and Geopandas processing 40+ million data points across the Greenland ice sheet.
  • Reduced analysis time from 3 minutes to 20 seconds per region using optimized interpolation techniques and Shapely geospatial processing.
  • Optimized ice sheet modeling software using Numba JIT compilation and NVIDIA CUDA parallel processing.
  • Created flexible data pipelines supporting various file formats with automated workflows for scientists.

Trauma THOMPSON - HoloLens AI Copilot

November 2024 – Present
Computer VisionMixed RealityHealthcare AIHoloLensActive Learning
  • Built YOLOv11 computer vision models for medical tool detection with 96% mAP@[0.5:0.95] using quantized models and focal loss.
  • Slashed data annotation needs by 80% with active learning pipeline using uncertainty sampling (PyTorch + Ray for distributed processing).
  • Crafted mixed reality system for HoloLens that visualizes ML inferences in real-time for first responders.
  • Contributed statistical analysis to research papers on AI-assisted emergency medicine.