Portfolio
All Projects
From agentic AI pipelines to computer vision systems — every project here is a real, shipped solution built to solve actual problems.
Agentic AI
FHIRFlow
Healthcare Claims Multi-Agent System
An end-to-end AI-powered pipeline for automating CPT/HCPCS medical claim validation. The system detects policy changes, updates patient records via FHIR, validates claims against live payer policies, routes approved claims for EDI 837 submission, and conducts real AI voice calls with patients for rejected claims.
💡 Automates end-to-end insurance claims — from patient visit to payer payment
Agentic AI
PR Reviewer Agent
Autonomous PR Review & CI/CD Deployment Agent
An autonomous AI agent designed to streamline development workflows by acting as a virtual team member on GitHub. It automatically intercepts pull requests, reviews code diffs for bugs, architectural design violations, and styling rules, runs test suites, generates code patches, and automatically redeploys CI/CD pipelines.
💡 Reviews pull requests, runs automated test suites, and auto-deploys CI/CD pipelines with bug patches
Generative AI
NEXUS AI / NeoVerse
Drug Trafficking Detection Multi-Agent System
An AI-powered multi-agent system that monitors platforms like Telegram, Reddit, and Instagram in real time to detect and classify drug-related content. Uses NLP and HDBSCAN clustering to identify drug trafficking discussion patterns.
💡 Real-time cross-platform drug trafficking detection at scale
NLP + Security
CVEE / NVDNLP
Autonomous Vulnerability Enforcement Agent
A local desktop security agent that continuously monitors known vulnerabilities on a system from the National Vulnerability Database and enforces remediation actions automatically — no alerts, no dashboards, no human delay. Uses NLP to analyze CVE severity scores and automatically stops affected services.
💡 Eliminates human delay in vulnerability response — monitors and remediates in real time
Computer Vision
MLOps Pothole Detection
Computer Vision + MLOps Pipeline
An AI-based computer vision project that detects potholes from road images or video using deep learning, with complete MLOps pipelines for model training, versioning, CI/CD, and scalable deployment for automated road damage monitoring.
💡 Scalable automated road damage monitoring with full MLOps lifecycle
Computer Vision
Weapon Detection Agent
AI Surveillance with Intelligent Decision Making
An AI-powered surveillance system that identifies weapons and potential threats in real time using YOLOv11 for object detection and Llama 3 as an intelligent decision-making layer to classify threat level and trigger appropriate responses.
💡 Real-time threat classification with intelligent escalation logic
Computer Vision
Polyp Segmentation
Medical Imaging Deep Learning
Uses deep learning to automatically detect and segment polyps in colonoscopy medical images, helping improve early detection and diagnosis accuracy.
💡 Improved early polyp detection for clinical diagnosis support
Community
Literate Spork
Open-Source GitHub Learning Community
A GitHub-based learning community that mentors 50+ students through 20+ hands-on projects in Data Science and Generative AI, fostering real-world skills through open-source, project-driven learning.
💡 50+ students mentored · 20+ hands-on projects · 4 stars on GitHub
ML & DL
DNASeq MLOps
DNA Sequence Classification Pipeline
An MLOps-powered pipeline that analyzes DNA sequences to classify whether a given sequence belongs to bacteria or not. Includes data preprocessing, feature extraction from genomic sequences, model training, and automated deployment.
💡 Automates biological sequence classification with scalable MLOps pipelines
ML & DL
Kidney Disease Prediction
Deep Learning Medical Diagnosis System
A deep learning-based system that predicts kidney disease using CNN models with high accuracy (~99%). Built with a complete MLOps pipeline including training, evaluation, and deployment for real-world clinical usage.
💡 High-accuracy medical diagnosis support system using deep learning
ML & DL
CodeCure AI
Molecular Toxicity Prediction System
An MLOps-based molecular toxicity prediction system that identifies toxic properties in chemical compounds. Uses explainable AI (SHAP) to provide feature-level contribution insights for each prediction.
💡 Improves drug safety analysis with explainable toxicity prediction