Mukund Biradar-AI Engineer | Sr Python Backend Specialist | Agentic AI | LLM Systems & RAG Pipelines
Check rate
Experience
Voice AI Chatbot - Real-Time Audio Assistant
- ▶ Built real-time voice assistant (STT → LLM → TTS pipeline) benchmarking and evaluating multiple STT providers including faster-whisper and Azure Speech. achieved sub-3s latency, Groq API (Llama 3) with multi-turn memory - directly handling edge cases in dictation, names and passcode recognition.
Senior AI Engineer / LLM Developer
Deloitte
- Architect and ship multi-agent AI workflows on Azure AI Foundry using a planner → manager → domain-agent hierarchy for enterprise automation across Finance Agent, Comparison agent and Operations personas.
- Specialized in building AI-powered customer-facing interfaces and conversational chatbots ,integrating LLMs, RAG architectures, agentic pipelines, Celeryworker scaling, Pydantic, type annotations, Azure Blob Storage, RabbitMQ and backend APIs to deliver end-to-end digital solutions for enterprise clients.
- Achieved 40% improvement in retrieval accuracy
- Built and owned production-grade versioned public-facing FastAPI APIs (v1/v2) with OpenAPI/Swagger documentation, rate limiting and sub-100ms latency - integrating LLM agents, event triggers and vector databases for semantic document retrieval.
- Achieved 70% database load reduction via Redis caching, query optimisation and horizontal scaling — maintaining 99.9% SLA on data services.
- Implemented OAuth2/JWT authentication, role-based access control (RBAC), Redis caching and pytest automation. deployed containerised services via Docker and CI/CD pipelines on Azure/AWS.
- Generative AI & LLM Built multi-agent LLM workflows on Azure AI Foundry using LangChain, RAG pipelines and vector databases for enterprise document retrieval. Delivered conversational AI interfaces with prompt engineering and LLMOps practices, achieving 40% improvement in retrieval accuracy.
- Designed and operated distributed Celery task queues with Redis and RabbitMQ brokers for asynchronous background processing, worker scaling and task retry strategies across multi-tenant AI workflows.
- Engineered end-to-end LLM solutions across the full stack Python, LangChain, LangGraph, Azure OpenAI, RAG, FastAPI, Docker and Kubernetes from embedding pipelines to deployed agentic workflows.
PDF-RAG LLM System
- ▶ Built end-to-end RAG system: PDF ingestion → text extraction → semantic chunking → vector embeddings → natural language Q&A over custom knowledge bases.
- ▶ Full LLM stack with versioned REST API layer (FastAPI + OpenAPI/Swagger), LangChain orchestration, AWS-compatible vector storage - designed for developer integration and external consumption.
Senior Python Developer
Leibniz-Rechenzentrum (LRZ)
- Designed and maintained high-throughput ETL/ELT data pipelines for structured and unstructured data ingestion, transformation and validation across PostgreSQL and AWS S3.
- Managed 5TB+ of research data on S3-compatible object storage (Minio/AWS S3), maintaining 99.9% availability with automated pipeline monitoring and alerting.
- Ansible: Deployed and managed Ansible playbooks to automate log rotation collection across 20+ distributed research servers, eliminating manual SSH-based log retrieval and enabling centralised log analysis - reducing ops overhead by 60%.
- Used Ansible to orchestrate package updates and configuration changes across the server fleet, ensuring consistent environment state and reducing configuration drift.
- Consulted on deployment and staging strategies (Dev → Staging → Prod), defining environment parity standards and rollback procedures for containerised AI services.
- Developed containerised Python microservices using FastAPI and asyncio, deployed via Docker for scalable backend systems.
- ↓ Reduced 70% database load reduction
- Led TDD culture with pytest and Robot Framework - full unit, integration and regression test suites for REST APIs and ETL pipelines.
Python Lead
IBM
- Led Python development teams delivering backend solutions and REST API integrations; authored architectural design documents and technical specifications.
- ↑ Built 80% reduction in manual effort
- Set up CI/CD pipelines with GitHub Actions for automated build and deployment; implemented OAuth2 and JWT authentication for secure API access.
- Containerised and orchestrated services using Docker and Kubernetes.
Python Consultant (Search & Data)
Intel
- Configured Coveo Search integrating XML, databases, Salesforce and AWS S3 for content indexing and retrieval pipelines processing 500,000+ records monthly.
- Triaged and resolved production defects, performing root-cause analysis across backend, database and integration layers.
- ↑ Achieved 40% improvement in search accuracy
Senior Software Engineer
Sinomonitor International
- Led full-lifecycle development of a teaching platform and e-commerce site using Django and PostgreSQL.
- Created and secured APIs serving internal teams and third parties, backed by multiple RDBMS backends (PostgreSQL, Oracle, MS SQL Server).
- Collaborated with database engineering teams to align on schema changes, stored procedures and data quality standards.
- Built ETL pipelines using Pandas and NumPy for data preprocessing, feature engineering and ML-based video analytics across 5+ source systems (Excel, MySQL, Oracle 11G/12 C, SQL Server).
- Designed scalable ETL architecture reducing reporting time from hours to minutes; created Tableau and Power BI dashboards for business intelligence.
Database Administrator
Develop Dreamz Industries
- Managed Oracle and MS SQL Server databases - query optimisation, schema design and data integrity across enterprise systems.
- Designed and developed application modules based on functional specifications and architectural blueprints.
Database Engineer
Psystems IT Services
- Designed and maintained database schemas, stored procedures and ETL scripts for data migration and reporting.
- Managed Oracle 11 G R1/R2 /12 C/ RAC and Data Guard environments, overseeing Clusterware, RMAN backup/recovery, DB Tuning and tablespace optimization to ensure enterprise-grade database stability and high availability.
DocsAI – Production-Ready Document RAG System
- Architected an end-to-end local RAG pipeline using FastAPI and React to perform semantic indexing and context-grounded natural language Q&A over dense PDFs and text files.
- Implemented sentence-aware text chunking paired with local 'all-MiniLM-L6-v2' embeddings and self-hosted ChromaDB vector storage inside Docker volumes to eliminate external cloud costs.
- Integrated the Groq API (Llama 3) using typed system guardrails to enforce source citation accuracy and mitigate LLM hallucination risks.
- Configured containerized multi-service deployment with Docker Compose and designed an enterprise production migration roadmap (AWS S3, pgvector/RDS, ECS Fargate, and ALB).
Industry Experience
See where this freelancer has spent most of their professional time.
Experienced in Information Technology, Education, Retail, Manufacturing, and Professional Services.
Business Area Experience
See which departments and functions this freelancer has contributed to most.
Experienced in Information Technology, Product Development, Business Intelligence, Operations, and Quality Assurance.
Summary
Senior AI Engineer and Python Backend Specialist with 12+ years building production systems at scale. Currently engineering multi-agent AI workflows and RAG pipelines for Deloitte on Azure AI Foundry, achieving 40% improvement in retrieval accuracy and 70% reduction in manual processing time. Strong foundation in distributed microservices, ETL pipelines, cloud-native deployments (Azure, AWS) and infrastructure automation with Ansible. Experienced across the full ML/LLM stack - from embedding pipelines and vector databases to agent orchestration and LLMOps with a track record of shipping measurable business impact in enterprise environments.
Skills
- Languages: Python (Primary), Java, Typescript, Javascript, Node.Js, Sql, Pl/Sql
- Ai / Llm: Langchain, Langgraph, Azure Openai, Rag Pipelines, Prompt Engineering, Llmops, Vector Databases, Semantic Search, Multi-Agent Orchestration, Nlp, Speech-To-Text (Stt), Whisper, Audio Processing
- Cloud & Devops: Azure Ai Foundry, Azure Devops, Aws, Docker, Kubernetes, Ansible, Aks (Azure Kubernetes Service), Docker Compose, Ci/Cd, Google Cloud Platform (Gcp)
- Frameworks: Fastapi, Django, Flask, React, Sqlalchemy, Sqlmodel, Asyncio (Async/Await), Type Annotations, Pydantic
- Agentic Dev Tools: Claude Code, Cursor, Github Copilot
- Databases: Cosmosdb, Postgresql, Oracle, Ms Sql Server, Redis, Minio S3
- Data Eng.: Etl/Elt Pipelines, Pandas, Numpy, Hadoop, Reverse Engineering
- Apis & Sec.: Restful Apis, Public Api Design, Api Versioning, Openapi/Swagger, Rate Limiting, Microservices, Oauth2, Jwt, Postman
- Testing: Tdd, Pytest, Robot Framework, Unit / Integration / Test Plan Authoring, Selenium
- Automation: Ansible, Shell Scripting
- Monitoring: Logging, Observability, Performance Monitoring, Debugging
- Bi & Reporting: Tableau, Power Bi, Deepinsight, Interactive Data Visualization, Event Tracking, Web Analytics
- Methodology: Agile (Scrum/Kanban), Jira, Azure Boards, Technical Documentation
Languages
Education
University of Pune
Master of Computer Science · Computer Science · India
Certifications & licenses
ChatGPT Masters: AI Prompt Engineering - 16 Hours
Azure AI Foundry - hands-on enterprise deployments
Machine Learning: Generative AI, Data Analysis with Pandas
Statistics
Experience
Global Experience
Expertise
Qualifications
Profile
Frequently asked questions
Have questions? Find more information here.
Similar Freelancers
Discover other experts with similar qualifications and experience
Experts recently working on similar projects
Freelancers with hands-on experience in comparable project as a Voice AI Chatbot - Real-Time Audio Assistant
