🚀 Passionate about building robust, end-to-end full-stack systems, scalable backend architectures, and production-grade AI/LLM pipelines.
📬 Reach out to me:
- 💼 Current Role: Associate Software Engineer at Oryx International Logistics (working on enterprise backend microservices, REST APIs, and database optimizations).
- 🎓 Education: B.Tech in Electronics & Communication Engineering (2025).
- 🧩 DSA Problem Solving: 290+ problems solved in C++ on LeetCode.
- 🛠️ Core Focus: Building high-performance RAG pipelines, secure multi-tenant architectures, and real-time streaming interfaces.
| Category | Technologies |
|---|---|
| Languages |      |
| Frontend |      |
| Backend |      |
| AI & LLM |  RAG Pipelines |
| Databases |     |
| DevOps & Cloud |     |
🧠 PanScience Q&A — AI Document & Multimedia Intelligence
Full-stack RAG platform enabling interactive Q&A over PDFs, audio, and video files.
- Backend: FastAPI with LangChain and pgvector for cosine similarity retrieval, integrated with OpenAI GPT-4o.
- Key Features: Server-Sent Events (SSE) for streaming text, Whisper API for timestamp-level citations, Redis for distributed rate limiting (SlowAPI) and chat history session management.
- Quality Assurance: Containerized using Docker Compose with 95%+ unit and integration test coverage.
💸 SpendSight — AI Spend Audit Platform
Live fintech-style platform auditing startup SaaS spend on AI tools. Live at spentsight.vercel.app.
- Stack: Next.js 14, TypeScript, Anthropic Claude API, Supabase (PostgreSQL).
- Security First: Implemented Supabase Row Level Security (RLS) policies for complete tenant data isolation.
- Deterministic Compute: Custom-engineered a TypeScript
AuditEnginewith zero LLM dependency to ensure flawless, deterministic financial calculations.
🚛 LogiTrace — Logistics Middleware & Analytics Dashboard
High-performance middleware simulating production workloads and dashboard analytics.
- Performance: Achieved a 35%+ database query latency reduction (benchmarked from ~200ms down to ~130ms) using optimized MongoDB aggregation pipelines.
- Type Safety: Shared Zod schemas between Next.js frontend and Express backend to enforce end-to-end runtime contract security.
- Simulation: Designed a custom Prisma-seeded data simulator generating hundreds of mock shipments across variable latency classes.
⚙️ AI-Powered Transaction Pipeline — Async Data Ingestion
Asynchronous transaction processing pipeline decoupling heavy computation from client endpoints.
- Architecture: FastAPI, Celery, and Redis backend, returning immediate job status tokens and polling asynchronously.
- Pipeline: 5-stage Celery worker flow integrating Google Gemini 1.5 Flash batching (20 records/call) with exponential backoff retries.
- Analytics: Statistical outlier flagging (>3x per-account median) combined with LLM anomaly classification.
I thrive on solving performance bottlenecks, designing clean databases, and engineering robust backend systems. If you're looking for a proactive software engineer who can jump into a complex codebase and ship high-impact features with high test coverage, let's connect!