Projects

1. Built a personal portfolio with Next.js App Router, TypeScript, and Tailwind CSS to showcase projects, experience, skills, and certifications.

2. Designed a modular, section‑based layout with reusable UI components and scroll‑triggered reveal animations.

3. Set up a GitHub Actions CI/CD pipeline to build the app and deploy the static export to cPanel via FTP on each main push.

1. Built a RAG-based AI system combining Code Property Graphs (CPGs) and LLM reasoning to detect and mitigate security vulnerabilities in C/C++ code.

2. Engineered a multi-agent architecture, where specialized agents analyze distinct vulnerability classes and resolve conflicting signals on shared code paths.

3. Integrated graph ML (PyTorch, DGL) with vector search (PineconeDB) and Gemini Pro to deliver context-aware, actionable security fixes.

1. Fetch and process NVD CVE 2.0 JSON feeds, handling schema validation, deduplication, and versioned output generation.

2. Normalise the data by extracting and structuring key fields such as CVE ID, description, severity scores, affected products, and published dates.

3. Mapped CVE records by products and their vendors using metadata from the CPE dictionary.

1. Architected event-driven backend systems using Kafka, NestJS, and GraphQL to replace legacy batch workflows and enable real-time inventory accuracy at scale.

2. Designed demand-based ordering and delivery-slot orchestration, automating purchase order generation while reducing unnecessary stock holding.

3. Engineered extensible product configuration schemas supporting complex variants and options without breaking existing APIs or impacting latency.

1. Built a RAG-based AI system to automate ABET accreditation document generation using LLMs and vector search.

2. Mapped student assessment data to ABET outcomes (1–7) through automated reasoning pipelines.

3. Reduced faculty workload by replacing manual compliance drafting with scalable AI-driven workflows.

1. Improved React Native performance on low-end devices by debugging JS thread contention and optimizing list rendering with FlashList.

2. Ensured low-latency mobile API responses through backend coordination and response optimization under unreliable network conditions.

3. Maintained backward compatibility across mobile releases by enforcing API versioning and non-breaking schema evolution.