Prathmesh Sadake

Prathmesh Sadake

AI Engineer & SDE · Mumbai, India · Open to relocating to Bangalore or Hyderabad

I build production AI systems that ship.

Generative AI products and scalable backends with Java, Spring Boot, Python, AWS, and Azure. Lately: cutting LLM costs by 60% for a major US solar provider, and email costs by 87% for my team.

View my workGet in touch

About

I'm a software engineer from Mumbai, India. I build AI-powered applications and the backend systems behind them.

At Build Fast with AI, I build production Generative AI systems on GPT, Claude, Mistral, LLaMA, and Gemini. I built a document processing pipeline on Redis queues and LLMs that has processed thousands of documents and cut manual work by 60%. Token optimization on a product for one of the largest solar providers in the US brought its LLM costs down by 60%. I also designed an internal email platform on AWS that cut email costs by 87%, and I've worked with our deployment team to ship applications to production on Azure.

Outside my day job, I build distributed systems in Java and Spring Boot with microservices, Kafka, Neo4j, Docker, and Kubernetes on GKE. I've also shipped 7 AI and web products for freelance clients and completed 6 Udacity Nanodegree programs.

I care about clean system design, measurable results, and helping the engineers around me grow.

Skills

AI / ML

LLM applications with GPT, Claude, Gemini, Mistral, and LLaMA · RAG · Multi-agent systems · MCP · Claude Agent SDK · LangChain · Token optimization

Backend

Java · Spring Boot · Python · TypeScript · Microservices · Kafka · Redis · Node.js · Next.js

Cloud & Data

AWS · Azure · Docker · Kubernetes · PostgreSQL · Neo4j · Supabase · Vector databases

Credentials

  • AWS Certified Solutions Architect · Amazon Web Services
  • 6 Udacity Nanodegrees · Agentic AI, AI Engineering with Claude, AWS Cloud DevOps, and more
  • Spring Boot 0 to 100 · Coding Shuttle, 2025

Experience

Build Fast with AI

Software Engineer 1, Generative AI

Present

  • Built a production-grade document processing pipeline using Redis-backed queues and LLMs to extract and process data, cutting manual workload by 60%.
  • Reduced LLM costs by 60% through token optimization on a tech product for one of the largest solar providers in the US.
  • Designed and shipped an internal email platform on AWS SES, SQS, Next.js, and PostgreSQL in 4 weeks, cutting email costs by 87%.
  • Built GenAI apps and custom pipelines for text generation, summarization, code assistance, image generation, and chat, using GPT, Claude, Mistral, LLaMA, and Gemini through OpenRouter.
  • Worked with the deployment team to ship applications to production on Azure.
  • Maintain backend services on PostgreSQL, Supabase, and vector databases, and mentor junior developers.
OpenAIAnthropic ClaudeAWSRedisPostgreSQLSupabaseNext.jsNode.jsPythonReactMySQLJavaScript

Software Development Intern

Aug 2023 to Jan 2024

  • Designed the database schema, relationships, and CRUD operations, and built a custom authentication system on Supabase.
SupabasePostgreSQLTypeScriptNext.jsReact

Bharat One Care

Software Development Intern

Oct 2023 to Dec 2023

  • Built RESTful APIs and Firebase authentication with role-based access control, and took part in code reviews.
FirebaseNode.jsJavaScriptReactMySQL

Ensemble Inc

Product Engineer

Nov 2022 to Oct 2023

  • Built responsive web apps with React.js and Next.js in an Agile team.
ReactNext.jsTypeScriptJavaScriptNode.js

TCS iON

Project Intern

Jun 2022 to Jul 2022

  • Built a deep learning model to detect emotions in English text and automate sentiment analysis of feedback.
PythonPyTorchpandasNumPyJupyter

Earlier roles: MERN Stack Developer at Docsup in 2022, and Web Software Developer at Arca Innovation in 2021.

Projects

AI Document Processing Pipeline

Queue-based LLM extraction that turns unstructured documents into structured data at production scale. Thousands processed, manual work down 60%.

Redis · LLMs · Azure

Distributed LinkedIn Clone

A social platform built as Java and Spring Boot microservices with a Neo4j connection graph and Kafka notifications, deployed on GKE behind an API gateway.

Java · Spring Boot · Neo4j · Kafka · Kubernetes

Custom Email Platform on AWS

Case study

Replaced a paid email provider with an SES and SQS pipeline with retries, bounce handling, and an admin dashboard. Shipped solo in 4 weeks, costs down 87%.

AWS SES · SQS · Next.js · PostgreSQL

Talkerz: AI English Speaking Coach

AI voice interviews that score fluency, pronunciation, grammar, and vocabulary in real time. 10K+ learners, 4.9 rating.

Next.js · Speech APIs · OpenAI · Firebase

Hirrings: AI Recruitment Platform

Built from scratch: AI candidate ranking, skill analysis against each job description, automated application tracking.

Next.js · Node.js · PostgreSQL · OpenAI

Agentic AI Projects

A multi-agent code review orchestrator, an MCP-powered pricing analyst, and UdaPlay, a RAG research agent with tool use and memory.

Python · Claude Agent SDK · MCP · RAG

More client work on the freelance projects page.