MADHAN MOHAN REDDY P
- |-| GitHub | Linkedin | Portfolio | Hugging Face
PROFESSIONAL SUMMARY
With 3.8+ years of experience in software development and AI/ML, I specialize in designing and
delivering Generative AI driven enterprise solutions, most notably architecting Agentic AI
systems by integrating the Model Context Protocol(MCP) as modular tools/plugins and
orchestrating complex workflows with the AgenticAI framework for maximum extensibility and
maintainability. I excel at optimizing knowledge retrieval, query processing, and response
generation for scalable, cloud-based AI platforms, and have deep expertise in building and
maintaining data/model pipelines using leading LLM and ML frameworks such as LangChain,
LangGraph, Semantic Kernel, Vector Databases, Scikit-Learn, TensorFlow, and PyTorch. I’ve
deployed production-grade Generative AI and Computer Vision solutions on Azure Cloud
leveraging Azure OpenAI, Azure Cognitive Services, Function Apps, and Container Registry,
while driving automation, system integration, and SQL-based data workflows. Currently
pursuing a PG Diploma in Machine Learning and Artificial Intelligence from IIIT Bangalore.
SKILLS
GenAI Frameworks: LangGraph · LangChain · Semantic Kernel · Advanced RAG · Azure
OpenAI· LlamaIndex · Vector Databases · Prompt Engineering · MCP (Model Context Protocol)
LLM Models : Azure OpenAI Models · OpenAI (GPT-3/4) · Ollama · Google Gemini · Hugging
Face (open-source) · Groq Models · Vision-enabled Models
Machine Learning: TensorFlow · Keras · PyTorch · Scikit-Learn · Regression Models · NLP
(NLTK, spaCy) · Data Cleaning · Data Analysis · Exploratory Data Analysis · Pandas · NumPy ·
Matplotlib
Cloud Services: Azure · AWS · Azure Cognitive Search · Azure AI Search · Azure Container
Registry · Azure Redis Cache · Azure Function & Web Apps
Programming Languages: Python · Java · SQL
DevOps & Deployment: Azure · Docker · Git
WORK EXPERIENCE
AI/ML Engineer at Xpheno - EY(Client)
Jan 2025 - Present
● Engineered and delivered Agentic AI applications built with Semantic Kernel and
LangGraph; integrated multiple LLMs (OpenAI, Azure OpenAI, Ollama, Gemini) and
●
●
●
●
●
●
leveraged the Model Context Protocol (MCP) as modular plugins to enable distributed,
client-specific AI workflows.
Developed and deployed Generative AI enterprise solutions using Azure OpenAI
models, Azure AI Search, and Redis, optimizing knowledge retrieval and response
generation.
Managed Function Apps for query response handling, embedding generation, and
document synchronization, ensuring efficient execution within a distributed cloud
environment.
Handled Azure SQL Database, Service Bus, Storage Accounts, and Azure File
Share, enabling secure and scalable data management.
Implemented containerized deployment strategies with Azure Container Registry
(ACR) to ensure seamless scaling and high availability.
Built and maintained CI/CD pipelines in Azure DevOps, automating build and
deployment processes for AI-powered applications.
Worked on backend development and optimized LLM-based query processing
pipelines, enhancing system performance, accuracy, and response efficiency.
Software Engineer at FIS GLOBAL (Fidelity Information Services) - Bangalore
Oct 2021 - Jan 2025
● Led the development of a Treasury Chatbot, a Retrieval-Augmented Generation (RAG)
application, integrating LLMs and vector databases for efficient financial data retrieval,
query handling, and intelligent decision-making in treasury operations.
● Built machine learning pipelines for automated data processing, feature extraction,
and model deployment, achieving 90% accuracy in predictive models.
● Enhanced banking applications by implementing new functionalities and fixing bugs,
ensuring compliance with US regulatory requirements and addressing client-reported
issues.
● Resolved defects and collaborated with BA and QA teams for North American banking
clients, contributing to core banking projects like US TAX forms and retirement
plans.
● Developed an automation project for the build process, improving project completion
speed by 35%.
● Created innovative prototypes using Generative AI and LLM models as part of FIS’s
Innovation Event
EDUCATION
PG Diploma in Machine Learning and Artificial Intelligence
International Institute of Information Technology, Bangalore(IIIT B) | 2025
Bachelor of Technology in Computer Science and Engineering (CSE)
Sir CV Raman Institute of Technology and Science, Tadipatri, Andhra Pradesh | 2021
Diploma in Computer Engineering (CME)
Government Polytechnic College, Proddatur, Andhra Pradesh | 2018
PROJECTS
● Uni-IDP: Developed an end-to-end Agentic AI solution for extracting key fields from
invoices and claims. Utilized Azure Form Recognizer OCR and Azure OpenAI GPT
models for precise text recognition and classification. Developed modular agents,
including Extractor, Planner, Validation, and Orchestrator, that leverage tools deployed
on MCP servers, and integrated Microsoft Teams as a Human-in-the-Loop bridge to
resolve ambiguous or missing data. Deployed on Azure Function Apps, Service Bus,
and Storage Accounts to ensure scalable, distributed processing and reliable document
synchronization.
● EY-Jarvis : EY-Jarvis is an AI-powered enterprise solution designed to enhance
knowledge retrieval and response generation. As part of the team, I’ve developed AI
modules using Azure OpenAI, Azure AI Search, and Redis cache, optimizing query
processing and system efficiency. I’ve managed Function Apps for query handling,
embedding generation, and document synchronization while ensuring smooth execution
in a distributed cloud environment. Additionally, we implemented CI/CD pipelines in
Azure DevOps and containerized deployment strategies with Azure Container Registry
(ACR) to enable seamless scaling and high availability. My contributions in backend
development and LLM-based query processing optimization significantly improved
system performance, accuracy, and response efficiency.
● SmartChainAgents(LLM Agent Routing System): Designed an intelligent query routing
system using LangChain and Groq's Gemma2-9b-It. Contextually routed queries to
appropriate sources such as Arxiv, Wikipedia, or an LLM, enhancing query handling for
diverse knowledge domains [GitHub].
● AIO GPT(Custom chat-bot): a custom chatbot capable of generating text, analyzing and
creating images, and summarizing YouTube videos. To ensure the chatbot remains
accurate and up-to-date, I created a Python package that fetches real-time information
and provides it to the LLM, ensuring the response is accurate [GitHub].
● SmartPDF(RAG): Developed a retrieval-augmented generation (RAG) application using
the Gemini-pro model for interacting with PDF documents. Integrated FAISS vector
search for efficient document indexing and retrieval. Built using Streamlit for an intuitive
user interface [GitHub].
ACHIEVEMENTS
● Awarded by FIS management for reaching the finals in the INNOVATE48 event and
recognized for outstanding contributions to automation.
