I build scalable full-stack systems and solve complex problems using C++ and modern web technologies.

• 🔭 Engineering full-stack platforms while mastering Data Structures & Algorithms in C++
• 🌱 Working with Next.js, TypeScript, and Serverless architectures
• ⚡ Building real-world systems: AI debugging tools & data-driven platforms
• 💬 Ask me about C++, Next.js, LLM Integration, or System Design

• CodeAutopsy → AI-powered debugging engine for dependency-aware root cause analysis

• OfferVault → Peer-to-peer placement registry bringing transparency to real, verified CTC data

The AI-Powered "Medical Examiner" for Broken Code

Tracing crashes in modern full-stack systems is a needle-in-a-haystack problem. Traditional debuggers show where it failed — not why across the full dependency graph.

Built an architecture-aware AI debugging engine that:

• Parses repositories into dependency graphs
• Traverses relationships using graph algorithms
• Applies LLM reasoning for contextual diagnosis

• Static Graph Analysis → Custom TypeScript parser builds a directed dependency graph and detects critical nodes

• Graph Traversal Engine → BFS with distance-penalty scoring to prioritize relevant context

• LLM Inference Layer → Gemini 3 Flash via Next.js API for multi-file reasoning

• Visual Debugging Layer → Mermaid.js diagrams highlighting the Critical Path of Failure

• Scalable Architecture → Upstash Redis (rate limiting) + Supabase (auth & data security)

• Performance Optimization → Repomix context packing + prompt tuning for high accuracy under token limits

• Model-Agnostic Design → Supports Groq, Gemini, Claude — easy provider switching

• ⚡ Automated root-cause detection across dependency chains
• 🧠 Faster onboarding for unfamiliar codebases
• 🏗️ Production-ready system with strict TypeScript & modern architecture

• 🧠 Advancing CodeAutopsy’s context-aware debugging intelligence
• ⚡ Scaling OfferVault for real-world peer usage
• 📈 Practicing Graphs, Dynamic Programming, and System Design

• 🔗 https://sidcore.in/ → Projects, deep-dives, and technical journey