PARTH GARG
Software Engineer
garg2006parth@gmail.com | www.parthgarg.me | candidates.intervues.club/u/parthgarg2006
SUMMARY
Software Engineer with experience in benchmarking, fine-tuning, and deploying LLMs and web applications. Proven track record of reducing model error rates and cutting inference costs for enterprise clients. Skilled in full-stack development, developer community mentorship, and backend engineering.
EXPERIENCE
LLM Intern
Pan Science Innovations LLP
Jan 2026 – Apr 2026
- Benchmarked 8+ LLMs and multiple ASR models for a major DTH client, driving a migration that cut inference costs by 30–40%.
- Engineered prompt strategies achieving a 25% reduction in model error rates.
- Fine-tuned and trained LLMs to meet client-specific requirements.
- Delivered GranthAI, an AI-powered learning platform for S Chand, including a production-grade PDF table extractor that resolved context loss in its RAG pipeline.
Web Development Head
IOSD MAIT
Jan 2025 – Present
- Mentored 50+ students in full-stack development, modern frameworks, and industry best practices.
- Organized and conducted workshops and hands-on sessions, building a strong developer community within the college.
EDUCATION
Maharaja Agrasen Institute of Technology · B.Tech · 2024 – 2028
MAIT Delhi · Current technical education and mentorship engagement · 2024
Prudence School · Class 12 CBSE · 2018 – 2024
SKILLS
Languages and Frameworks: C/C++, Go, JavaScript, Python, SQL, TypeScript, Bun, Echo, Express.js, FastAPI, Fastify, Flask, Next.js, Node.js, React.js, Streamlit, Zustand
Databases: MongoDB, MySQL, PostgreSQL, QdrantDB, Redis, VectorDB
Tools: AWS (EC2, S3), Cloudflare, Docker, Git, GitHub Actions, Jupyter, Postman, Prisma, RabbitMQ, Webhooks
PROJECTS
Zero-Devops: Self-Hosted Deployment Platform
Architected a 4-workspace monorepo (Next.js client, Go/Echo API server, Go build worker, shared schema contract) powering a self-hosted deployment platform. Designed a transaction system using Go, Echo, Next.js, PostgreSQL, and RabbitMQ.
Go, Echo, Next.js, PostgreSQL, RabbitMQ
Voltra: Low Tension Line Fault Detector
Developed a smart grid monitoring system using basic ML classification on SCADA data to predict LT line faults in real-time. Built a dual-interface system with admin dashboard using MERN Stack, Flask, Python, and Gemini LLM.
MERN Stack, Flask, Python, Gemini LLM
Zero-Devops
DevOps automation and infrastructure tooling designed for faster developer shipping and environment provisioning. Implemented using Go, Cloudflare, Echo, and CI-CD pipelines.
Go, DevOps, Backend, CI-CD, Cloudflare, Echo
Additional projects: Voltra - LT Fault Detection Service, LingoDocs
AWARDS
Competed in 20+ hackathons, GitHub Contributions, Class 12 CBSE Result