Applied AI Engineer · LLM Systems · Production AI · 14 years

Building AI systems that
think and scale.

I build production-grade AI pipelines — document intelligence, LLM integration, and RAG systems — backed by 14 years of backend engineering on AWS. Most engineers add AI to their stack. I build the part that makes it trustworthy.

Applied AI engineer based in Pune, India, with 14 years across travel, logistics, food-tech, and AI document processing at companies like American Express, FedEx, and Punchh.

At Atos, I built an AI-powered invoice reconciliation pipeline using AWS Bedrock, Textract, and Python — processing thousands of documents monthly at ~94% accuracy in production.

My edge: I don't just integrate AI — I build the evaluation, fallback, and reliability layer that makes it production-safe.

  • 14Years of production experience
  • 94%AI reconciliation accuracy in production
  • 5+Domains: travel, logistics, food-tech, AI

01

AI Invoice Reconciliation Pipeline

Production system processing thousands of invoices monthly. Multi-modal — AWS Textract for clean documents, Bedrock Vision for low-quality scans — with a custom LLM-powered matching engine at ~94% accuracy.

Python AWS Bedrock Textract LLM S3 Lambda Ruby on Rails

02

Ola Maps Ruby Gem

Open-source Ruby gem wrapping the Ola Maps API. Reposted by Ola's CEO — one of the first community-built integrations for the platform.

Ruby Open Source Maps API

03

Secrets Manager Reader Gem

Ruby gem wrapping AWS Secrets Manager for Rails-idiomatic credential retrieval. Centralises secret management across microservices with zero boilerplate.

Ruby AWS Secrets Manager DevOps

AI / ML

  • Python
  • AWS Bedrock
  • LLM Integration
  • RAG Pipelines
  • Prompt Engineering
  • OCR / Textract

Backend

  • Ruby on Rails
  • REST APIs
  • PostgreSQL
  • Redis
  • Sidekiq

Infrastructure

  • AWS (S3, Lambda)
  • Docker
  • CI/CD
  • Cloudflare
  • Git / GitHub
2022 — now

Senior Software Engineer · AI Systems

Atos

Built production AI invoice reconciliation pipeline using AWS Bedrock, Textract, Python, and Ruby on Rails. Processing thousands of documents monthly at ~94% accuracy.

2019 — 2022

Senior Backend Engineer

Punchh (PAX Technology)

Customer loyalty and engagement platform for restaurant chains. Built rewards, offers, and real-time infrastructure at scale.

2016 — 2019

Backend Engineer

FedEx / homify

Logistics and home design platforms. APIs, data pipelines, and third-party integrations.

2011 — 2016

Software Engineer

American Express Travel

Travel platform backend — booking, pricing, and itinerary management systems.

Available for global opportunities.

Targeting Applied AI and LLM engineering roles in Europe, Canada, and Australia. Open to remote with US or European companies.

hi@shobhitbuilds.com ↗