I'm an undergraduate Computer Science student at PES University specializing in Artificial Intelligence and Machine Learning. I build data-driven AI systems that solve real-world problems across healthcare, infrastructure monitoring, and autonomous systems. My focus is on developing explainable, production-grade ML pipelines with hands-on experience in sensor telemetry, time-series anomaly detection, and reinforcement learning.

• Autoencoders, LSTM, CNN
• Anomaly Detection, Time-Series Forecasting
• SHAP/XAI, Extreme Value Theory (SPOT/DSPOT)
• Sensor Data Analysis, Predictive Modelling
• Adaptive Thresholding, MLOps Fundamentals

Publication-ready | Hybrid Deep Learning + Explainability Pipeline | Repository: GitHub

• Engineered anomaly detection achieving F1 score 0.9959 and 16x improvement in precision using Variational Autoencoders + DSPOT adaptive thresholding
• Built latent-augmented XGBoost classifier with 96.85% test accuracy and F1 of 0.9676
• Proposed novel XAI Coherence Score (XCS) metric to measure cross-stage interpretability agreement
• Applied SHAP-based explainability to identify critical process parameters in fault classification

Tech: Variational Autoencoders, XGBoost, SHAP/TreeSHAP, Extreme Value Theory, Multivariate Sensor Telemetry

Production-Grade Multimodal AI System | Repository: GitHub

• Architected end-to-end AI system integrating biometric identity recognition (ArcFace face + ECAPA-TDNN voice) with Retrieval-Augmented Generation
• Developed vector-database-backed memory system with sub-second retrieval latency
• Full STT → LLM → TTS pipeline combining Whisper, Qwen LLM, and Coqui TTS for conversational prompts
• Demonstrated production-grade architecture design from scratch

Tech: ArcFace, ECAPA-TDNN, RAG, Qwen LLM, Whisper STT, Coqui TTS, ChromaDB

Live App: civic-spark-nine.vercel.app | Repository: GitHub

A modern, mobile-first civic engagement platform enabling citizens to report, vote on, and track civic infrastructure issues while earning XP, badges, and climbing leaderboards.

Key Features:

• Interactive OpenStreetMap-powered dashboard with real-time issue markers (color-coded by status)
• Gamification system: 5-tier progression (Citizen → City Champion), achievement badges, daily streaks
• Community voting with weighted user influence based on reputation levels
• Issue lifecycle tracking: Report → Verified → Acknowledged → In Progress → Resolved
• Trust & verification system with photo evidence, GPS timestamping, and spam prevention
• Smart priority calculation algorithm balancing upvotes, trust scores, and time decay

Impact: Designed for citizen-government collaboration in Bengaluru's civic infrastructure management

Tech Stack: React 18, TypeScript, Vite, Tailwind CSS, React Leaflet, Radix UI, Framer Motion, TanStack Query

Adaptive Deep RL Agent | Repository: GitHub

• Trained PPO-based deep RL agent in custom Gymnasium environment for real-time traffic optimization
• Engineered reward functions balancing waiting time, congestion, and emergency vehicle prioritization
• Implemented learning-rate annealing for stable generalization to unseen traffic scenarios

Tech: Gymnasium, Proximal Policy Optimization (PPO), Custom Simulation Environment

IoT + ML for Forest Fire & Landslide Detection | March 2025

Tech: Arduino, IoT Sensor Fusion, Anomaly Detection, Threshold Optimization

Motorsport Performance ML Pipeline | October 2025

Tech: Random Forest, Logistic Regression, XGBoost, FastF1 Telemetry, Streamlit

• Explainable AI (XAI) – Interpretability in deep learning models
• Reinforcement Learning – Sequential decision-making and adaptive systems
• Quantitative Machine Learning – Statistical foundations and time-series modeling
• Intelligent Transportation Systems – Autonomous vehicles and traffic optimization
• AI for Healthcare & Infrastructure – Real-world application of ML to critical systems

• Deep Reinforcement Learning for complex decision problems
• Advanced RAG architectures for knowledge-intensive AI systems
• Interested in Quantitative Machine Learning, applying statistical modeling and machine learning to financial data.

I'm interested in collaborating on AI research projects, open-source ML systems, and interdisciplinary applications of machine learning in healthcare, infrastructure, and autonomous systems.

If you're working on something interesting in the ML/AI space, feel free to connect!