AI Financial Intelligence Platform
AI-powered platform for financial data analysis, LLM-driven sentiment scoring, earnings call summarization, and automated investment insights.
Things I've built end to end, from data pipelines and models to deployed apps, and the decisions behind them.
AI-powered platform for financial data analysis, LLM-driven sentiment scoring, earnings call summarization, and automated investment insights.
Automated monitoring system for ML pipelines — tracks data drift, pipeline health, model performance metrics, and alerts on anomalies.
Predictive models for FIFA World Cup 2026 match outcomes using Elo ratings, historical team performance, and tournament simulations.
ML-based prediction model for Bengal state election outcomes using polling aggregation, historical voting patterns, and demographic data.
ML model predicting IPL 2026 cricket tournament outcomes using historical match data, player statistics, and ensemble methods.
Retrieval-Augmented Generation pipeline for querying documents with LLMs, backed by a NoSQL vector store for semantic search.
EDA and time-series forecasting of Delhi air quality data using statistical models and ML-based approaches for pollution trend prediction.
Query latency benchmarking across NoSQL (MongoDB, Redis) and SQL (PostgreSQL) — comparing read/write performance under varying workload patterns.
Interactive dashboard for monitoring model drift, data distribution shifts, and performance degradation in production ML systems.
NLP pipeline for analyzing sentiment in climate news articles using transformer-based models and VADER lexicon analysis.
Autonomous multi-agent system where specialized LLM agents collaborate to research topics, gather evidence, and produce synthesized reports.
Unified explainability framework for multiple ML models using SHAP, LIME, and attention-based visualization for model interpretability.
Distributed system of LLM agents that deliberate, debate, and reach consensus on complex queries through structured reasoning and voting.