Looking for a job

Rohan Mulay

@rohanmulay

Message

AI Engineer shipping production LLM and edge AI systems—building scalable agents, RAG, and MLOps pipelines.

India

Message

What I'm looking for

I’m looking for a role where I can build production LLM and agentic systems (RAG, evaluation, MLOps) and ship measurable outcomes with tight cross-functional collaboration—ideally including real-time or edge AI deployment.

I’m an AI Engineer focused on turning stakeholder requirements into production-grade LLM applications, agentic workflows, and scalable edge inference systems. I take end-to-end ownership—from architecture and prompt/model evaluation to MLOps practices that keep systems reliable in real environments.

In my current role, I’ve built an AI candidate screening platform with LangGraph, FastAPI, and Redis, using optimized prompt engineering and multi-model routing to evaluate 300+ profiles in parallel and cut screening time by 65%. I also led a RAG resume optimization platform (Next.js, Qdrant, SSE streaming) that reduced baseline hallucinations by 40%+ through structured model evaluation, and deployed 11 automation workflows that reduced manual GTM overhead by 80%.

I’ve extended this to real-time voice AI and on-device computer vision: I built low-latency STT/TTS voice agents with Vapi and Twilio (4.2+ MOS across 200+ calls), and for ADAS/edge AI I reduced inference latency from 42ms to 11ms using TensorRT FP16/INT8 on NVIDIA Jetson. Across projects, I prioritize measurable outcomes, model evaluation rigor, and systems that integrate cleanly with product and firmware teams.

Experience

Work history, roles, and key accomplishments

Current

AI Engineer Intern

Current

BeGig Studio

Apr 2026 - Present (3 months)

Owned end-to-end delivery of an AI candidate screening platform using LangGraph, FastAPI, and Redis with multi-model routing, evaluating 300+ profiles in parallel and reducing client screening time by 65%. Built additional systems including a RAG resume optimization platform (Next.js, Qdrant, SSE) and real-time voice AI agents using Vapi and Twilio with JWT authentication, reaching 4.2+ MOS averag

LangGraph fastAPI Redis Prompt Engineering Model Routing Model Evaluation RAG Next.js Qdrant Server Sent Events (SSE)Vector Search VAPI Twilio TTS JWT Workflow Orchestration

AI Engineer Intern

BYTES / Cautio

Oct 2025 - Apr 2026 (6 months)

Developed scalable ADAS perception and Driver Monitoring System features, including FCW and lane segmentation, validating precision/recall across firmware releases and deploying FastAPI microservices with vector search for real-time lookups. Reduced inference latency from 42ms to 11ms on NVIDIA Jetson Orin NX by deploying YOLOv11 with TensorRT FP16 and INT8 quantization while maintaining 91.3% mAP

Forward Collision Warning (FCW)YOLOv11 TensorRT INT8 Quantization fastAPI Vector Search Driver Monitoring Systems

ADAS & Machine Learning Engineer Intern

Moonrider.ai

Apr 2025 - Oct 2025 (6 months)

Owned an end-to-end semantic segmentation pipeline for drivable area detection, defining data architecture and using LiDAR-grounded augmentation across 15K+ frames to reach 83.4% mIoU. Improved edge deployment scalability using pruning and INT8 quantization (62% smaller model, 24 FPS) and built LSTM-based battery SoC/SoH prediction achieving 2.1% RMSE over 800+ charge cycles.

Pruning INT8 Quantization Battery State Of Charge (SoC)Battery State Of Health (SoH)