Headline

Software Engineer | Gen AI Engineer @Coditas | Agentic Systems, Voice Automation & Generative AI | Claude, LLMs, MCP, OpenAI

About

AI Engineer building production-grade systems at the intersection of generative AI, voice automation, and agentic workflows. I ship AI that thousands of users depend on daily — not prototypes, but systems that scale. At 64 Squares LLC, I built LLM-powered voice assistants, RAG pipelines, and multi-agent workflows that transformed property management operations: - Achieved 92% accuracy in real-time defect detection using YOLOv8 + CoreML on iOS at 30 FPS - Cut tenant query resolution time by 40% with RAG systems built on OpenAI + FAISS/ChromaDB - Reduced manual intervention through multi-agent automation using CrewAI and LangGraph At CloudBuilders Technologies, I engineered an intelligent voice routing system for an AI Voice Bot platform — enabling real-time, context-aware handoffs from automated agents to human reps. Stack: Twilio, Plivo, Pipecat, ElevenLabs, OpenAI, and Gemini. Currently at Coditas (Pune) as a Gen AI / Python Engineer, I build AI-powered workflow automation systems that dynamically trigger tools and extract structured data on demand. I use Claude Sonnet and Claude Opus with Windsurf to ship features fast and resolve complex bugs. Core skills: Python · LLM engineering · RAG · multi-agent systems (CrewAI, LangGraph) · computer vision (YOLOv8, CoreML) · voice AI (Twilio, ElevenLabs, Pipecat) · OpenAI · Claude API · FAISS · ChromaDB · agentic AI · workflow automation Certified in Claude 101, Claude Code 101, Claude Code in Action, Agent Skills, and MCP. Open to collaborating on generative AI, voice automation, agentic systems, and production ML projects. Connect with me or message me directly.

Location

Pune District

AI stack

anthropic apis

claude code

claude

cursor

elevenlabs

github copilot

hugging face

langchain

langgraph

langsmith

mcp

n8n

notion

openai

Work experience

Software Engineer | Claude, MCP, LangGraph, RAG, Python at Coditas, Jan 2026 - Present · 7 mos, Pune City, • Build an AI-powered workflow automation platform processing 100+ automations daily using Claude Sonnet and Claude Opus, achieving 92% data extraction accuracy via RAG-augmented prompt chains (up from 78%). • Architect Model Context Protocol (MCP) server integrations with 10+ custom domain-specific MCP servers, expanding platform capabilities 5x and enabling any MCP-compatible LLM to interface with the system. • Reduce workflow execution latency by 20% and token costs by 25% through intelligent LLM response caching and dynamic model selection across Claude, OpenAI, and Groq backends. • Ship 5+ major features (workflow composition, model switching, result export) with 95%+ test coverage and zero critical regressions; resolve 15+ LLM integration bugs improving response consistency by 18%.

AI Engineer | Voice AI, Pipecat, LLM, Twilio, ElevenLabs, OpenAI at CloudBuilders Technologies, Oct 2025 - Dec 2025 · 3 mos, Hyderabad, • Engineered a production-grade intelligent call routing system for an AI Voice Bot platform serving multiple enterprise clients, enabling real-time handoffs from automated agents to human representatives. • Designed and shipped a context-aware Transfer Call to Human Agent feature using conversation analysis and LLM-driven intent detection — reducing misrouted calls and improving resolution rates for live support teams. • Integrated Twilio, Plivo, Pipecat, ElevenLabs, OpenAI, and Gemini in a sprint-based agile workflow, delivering the feature end-to-end within a 2-month engagement. • Built real-time voice pipeline logic in FastAPI, handling concurrent call sessions with sub-500ms routing decisions.

AI/ML Engineer | LLMs, RAG, LangChain, CrewAI, YOLOv8, CoreML, Python at 64 Squares LLC, Mar 2024 - Oct 2025 · 1 yr 8 mos, Pune District, • Deployed YOLOv8-based defect detection on iOS via CoreML achieving 92% accuracy at 30 FPS, processing 100+ property inspections daily across 500+ properties with under 200ms inference latency. • Built a production RAG system (OpenAI + FAISS/ChromaDB) that cut tenant query resolution time by 40% and reduced API costs by 30%, serving 2,000+ monthly interactions at 95%+ response accuracy. • Architected multi-agent workflows with LangGraph and CrewAI managing 50+ concurrent agent instances, reducing manual real estate task intervention by 65% across automated pipelines. • Optimized FastAPI inference servers for 99.2% uptime and sub-500ms latency under 10,000+ concurrent users; built Snowflake/SQL data pipelines cutting model training time by 30%.

DevOps Intern | CI/CD Pipelines, Jenkins, Terraform, AWS at Tata Consultancy Services, May 2023 - Jul 2023 · 3 mos, Pune District, • Built and deployed CI/CD pipelines in Jenkins, reducing manual release effort and cutting integration errors across multiple deployment environments. • Provisioned and managed AWS infrastructure (S3, Lambda, EC2) using Terraform (IaC), establishing reusable modules that improved environment consistency across development and staging. • Integrated AWS Lambda and S3 with Jenkins workflows, automating storage and compute provisioning for each deployment cycle. • Enforced Git branching standards and code review workflows across a team of 5+ engineers, improving merge quality and CI pass rates. • Delivered CI/CD best-practices training to the team, reducing pipeline misconfiguration incidents during the internship period.

Education

Master of Computer Applications - MCA, Data Science, MIT College of Management - India

Bachelor's degree, Computer Science, Swami Vivekanand College Kolhapur

Public links

github: https://github.com/ShubhamChougale01