I build backend systems and AI tools under real-world constraints: limited hardware, noisy inputs, and users who value speed and clarity over complexity.

My work focuses on making language models and automation useful—especially on lightweight infrastructure and edge devices. I care about efficiency and robustness, not just benchmarks.

I value clean abstractions, measurable trade-offs, and software that survives contact with reality.

STEMist Education — Chapter Lead & Technical Lead. Founding the STEMist chapter at Boys' High School, Prayagraj; building a zero gate-keeping, builder-focused STEM club with hands-on workshops and demo days.

Bits&Bytes — Founding Member & Technical Lead. Shaping vision and running hackathons, coding workshops, and tech events for young developers; managing website, comms, and event systems.

Previously Jr. Research Engineer at jhana (Nov 2025–Jan 2026): AI-powered legal tech—WhatsApp conversational flows with Gemini voice transcription, multimodal legal consultation analysis, natural-language command parsing for document editing and version control, concurrent processing for real-time message streams.

Languages: Python, TypeScript, Dart, Bash, SQL
Backend & APIs: FastAPI, Next.js, REST APIs, LangChain, Discord.py, Telegram Bot API
AI / ML: LLMs, OpenAI API, Groq, fine-tuning, NLP, Docker, LLMOps
Hardware & edge: Raspberry Pi, ESP32, Arduino
Cloud & data: AWS, EC2, Azure, SQLite, Pinecone, Vercel
Tools: Linux, Git, GitHub, CI/CD basics

• Efficient inference and model compression (quantization, LoRA/QLoRA, lightweight fine-tuning)
• Retrieval-augmented generation for structured and semi-structured data
• Edge and on-device AI where latency, power, and privacy matter
• Systems that stay useful under pressure—robust when sensors and inputs are noisy