Headline

Backend & GenAI Engineer | Python • Node.js • PostgreSQL • LLMs | Ex-Reliance Jio Intern | VIT Vellore ’26

About

I don't just write code - I build systems that solve real problems. Last year at Reliance Jio, I built a multi-agent GenAI assistant from scratch for Telecom APIs. Before it existed, developers spent 2–3 hours resolving API queries. After? Under 1 minute. That's the kind of impact I chase. I'm a final-year B.Tech CSE student at VIT Vellore , specializing in backend engineering and AI/LLM systems. My stack: Backend: Python, Node.js, Express.js, Flask, REST APIs Databases: PostgreSQL, SQLAlchemy, query optimization AI/GenAI: Langflow, RAG pipelines, Qwen/Mistral LLMs, Langfuse, Prompt Engineering DevOps: Docker, Git, Render, Agile sprints Beyond my internship, I've independently shipped two live products: Personal Finance Tracker (Live) → https://fj-be-r2-yashsinghal-vitvellore-h9q3.onrender.com/ SQL Query Optimizer Pro (Live) → https://sql-optimizer.onrender.com/ I'm actively looking for Backend, GenAI, Data Engineering, or Full-Stack roles at product companies and MNCs. I'm open to relocation and remote opportunities. Let's build something great. 🔗 github.com/Yash19Singhal

Location

India

AI stack

antigravity

claude code

claude

codex

cursor

gemini

Work experience

Systems Engineer Trainee at Infosys Ltd, Aug 2026 - Present · 1 mo, Mysore

Software Intern at Reliance Jio, Jun 2025 - Jul 2025 · 1 mo, Navi Mumbai, Maharashtra, India, Built and deployed a production-grade multi-agent GenAI assistant for Telecom APIs, reducing developer query resolution time from 2–3 hours to under 1 minute while improving system throughput by 35%. • Designed a multi-agent AI pipeline (Orchestrator, Search, Analysis, Generation) handling ambiguous natural language queries for Telecom APIs — cutting resolution time from 2–3 hours to under 1 minute. • Integrated RAG search, OpenAPI parsing, and code generation into a modular orchestration workflow, reducing developer onboarding time by 40% and improving system throughput by 35%. • Achieved 99% accuracy in API schema extraction using semantic retrieval (Roadie RAG + OpenAPI parsing), validated across 20+ live queries from Jio Developer Studio. • Benchmarked Qwen 2.5 vs Qwen 3.14 on Ollama — reduced average response latency by 25% through model selection and prompt optimization. • Integrated Langfuse observability, cutting troubleshooting effort by 50% and standardizing prompt templates to reduce recurring errors by 25%. • Collaborated with backend, API, and product teams in Agile sprints — contributed to design reviews, sprint planning, and cross-functional delivery.

Education

Bachelor of Technology - BTech, Computer Science, Vellore Institute of Technology

Public links

github: https://github.com/Yash19Singhal