Open to opportunities

Ashish kumar Pati

@ashishkumarpati

Message

AI Engineer specializing in multi-agent LLM systems and scalable health analytics.

India

Message

What I'm looking for

I want to build production-grade multi-agent LLM systems, ship scalable AI inference on AWS, and keep costs efficient. I’m excited to own architecture end-to-end, deliver traceable RAG outputs, and collaborate with cross-functional teams.

I’m an AI Engineer with 8+ years of experience in backend development, focused on multi-agent LLM systems and health technology. I’m currently architecting GabitOne, a production-grade health analytics platform that ingests biomarker data from blood panels, wearables, and IoT devices using a 3-agent pipeline deployed on AWS EKS.

In my roles, I’ve built scalable AI inference systems and driven major efficiency gains—most notably reducing LLM inference costs by 70% through tiered model routing, prompt caching, batch processing, and replacing non-critical LLM calls with deterministic logic. I also deliver traceable, defensible outputs using a hybrid RAG approach that combines BM25 lexical search with vector embeddings for clinical evidence retrieval.

I enjoy leading cross-functional engineering teams and turning AI into reliable products. Previously, I led a 10-member team to ship AI-powered agricultural and school nutrition monitoring solutions, including edge AI modules and computer-vision-based monitoring systems that improved response times and operational accuracy.

Experience

Work history, roles, and key accomplishments

Current

AI Engineer

Current

Gabit

Jan 2025 - Present (1 year 4 months)

Solely built GabitOne, a multi-agent health analytics platform ingesting biomarker data (blood panels, wearables, smart scales) to generate personalized insights. Architected a 3-agent pipeline and deployed an async FastAPI service on AWS EKS with DynamoDB/SQS, implementing hybrid RAG and reducing LLM inference costs by 70%.

Hybrid RAG fastAPI DynamoDB Prompt Engineering

Technical Lead

AgNext Technologies

Jan 2021 - Jan 2025 (4 years)

Led a 10-member cross-functional team delivering AI-powered agricultural quality solutions and built VisioPrime, an edge computing module enabling real-time grain quality predictions with direct cloud connectivity. Developed an MDM monitoring system using OpenCV face recognition and sensors, and reduced server response time by 30% through performance optimization.

Edge Computing IoT OpenCV Computer Vision Python Django REST APIs Docker

Senior AI Engineer

Capillary Technologies

Jan 2018 - Jan 2021 (3 years)

Optimized retail merchandising using AI-powered customer analytics (age/gender detection), improving platform efficiency by 30%. Built computer-vision footfall counter systems to track customer movement and derive conversion and peak-hour insights.

Computer Vision AI & Analytics Python Express Microservices MongoDB Retail Analytics REST APIs