Looking for a job

Vijay Fagaria

@vijayfagaria

Message

SDE II at SpeEdLabs building RAG and AI agents, cutting LLM costs by about 40%.

India

Message

What I'm looking for

I’m interested in the part of AI engineering that begins where the demo ends: making systems reliable, efficient, and useful in the real world. I enjoy solving challenging engineering problems and turning ideas into production-ready systems.

At SpeEdLabs, I build AI features and the backend for an EdTech SaaS platform. I replaced whole-chapter prompts with a LangChain RAG pipeline, cutting tokens by over 90% and raising question-deduplication precision from 93% to 99%.

I also optimized LLM workflows across OpenAI, Anthropic, and Gemini, reducing inference costs by about 40%. For OMR grading, I used OpenCV and confidence scoring to improve scan accuracy from 80% to 98% and cut teacher review by 97%.

I shipped a LangGraph agent that prepares customer success managers for school visits with source-cited briefs, and moved grading and AI jobs onto Amazon SQS workers. I also designed pytest checks against real PostgreSQL and MongoDB for the CI/CD pipeline.

Previously, at Neuranics Lab, I developed Python software for AI-based diagnostic devices, including instrument control and image analysis. I deployed a cell-detection system at AIIMS that matched lab reports with 93% accuracy; I’m an IIT Kanpur graduate interested in making AI systems reliable and useful beyond the demo.

Experience

Work history, roles, and key accomplishments

Current

Software Development Engineer II

Current

SpeEdLabs

Nov 2025 - Present (11 months)

EdTech SaaS platform used by 450+ schools and 225K students. I build its AI features and the backend underneath them.
• Replaced whole-chapter prompts for question deduplication with a LangChain RAG pipeline that sends only the 5 closest matches from hybrid vector and keyword search, cutting tokens by over 90% and lifting precision from 93% to 99%.
• Optimized a multi-provider LLM pipeline across

Software Development Engineer

Neuranics Lab

Jun 2023 - Oct 2025 (2 years 4 months)

AI-based diagnostic devices.
• Developed a multithreaded Python instrument-control service, synchronizing camera, 3-axis stage motion, and image capture through a thread-safe command queue for stable unattended operation across long imaging runs.
• Introduced a config-driven hardware abstraction layer that enabled new camera backends without application code changes.
• Implemented cell-detection a

Software Development Intern

SmartTraveller

Jul 2022 - Jul 2022 (0 months)

Built a Node.js and Express recommendation service aggregating travel inventory from multiple provider APIs into MongoDB, scoring options on weighted price, duration, and rating signals for per-user ranking.