Karan Kadam

I’m an AI/ML Engineer with 4+ years of industry experience, specializing in AI, ML, NLP, Generative AI, LLMs, and Agentic AI. I design, build, and deploy end-to-end ML, LLM, and agentic systems across cloud environments, including model deployment and monitoring, to solve complex business problems and deliver scalable, high-impact solutions that drive real value.

Workers Compensation using Generative AI, Client - Verinext (09/2024 – 12/2025)
This legal-tech project focuses on streamlining the workers' compensation process through advanced Generative AI solutions. The aim is to automate and optimize the generation of critical legal documents such as Master Summaries, Pre-Hearing Memos, and Demand Strategies using the GPT-4o mini model. The project builds on our earlier legal case identification framework, enhancing it to support end-to-end legal content creation workflows. It leverages Salesforce Litify data, processed through a robust GenAI pipeline to provide precise, structured legal outputs for law firms.
Key Contributions:
• UI Development: Designed and implemented the frontend interface for user interaction and document generation.
• Data Integration: Built scalable data extraction pipelines from Salesforce Litify for ingesting case-related information.

• Prompt Engineering: Crafted and optimized prompts tailored for legal document  generation using GPT-4o mini. 

• Modeling: Integrated the LLM with RAG (Retrieval-Augmented Generation)  architecture using Langchain and FAISS-DB. 

• Testing: Developed unit test coverage for GenAI pipelines to ensure reliability  across different document types. 

• Deployment: Deployed the application stack (FastAPI + Flask) on Azure Virtual  Machines using Azure DevOps for CI/CD. 

• Technologies Used: GPT-4o mini, Langchain, RAG, FastAPI, Flask, FAISS-DB, Azure  OpenAI, Azure DevOps, Salesforce Litify, ADF (Azure Data Factory)

---------------------------------------------------------------------------------------------------------------------------------------------------------

Legal Case Identification Using Generative AI, Client- Verinext (07/2023 – 09/2024)

In this project, I worked for a law firm management company, where we aimed  todevelop an AI-powered system that could accurately identify various fields in  legalcases. The system was designed to streamline the process of identifying  relevantinformation from large volumes of text data, thereby improving the  efficiency andaccuracy of legal case analysis. The project involved utilizing cutting edge naturallanguage processing (NLP) techniques to extract insights from legal  casedescriptions. We employed a combination of technologies such as Langchain,  RAG,LLM, GPT3.5, FAISS-DB, Flask, FastAPI, Azure OpenAI, Litify, ADF to build a  robustand accurate model. 

Key Contributions: 

• Architected: Developing the architecture of the AI model using Langchain andRetrieval Augmented Generation.
• Implemented: Implementing the Azure OpenAI pipeline using GPT 3.5,Litify, ADF,FAISS-DB and FastAPI
• Trained: Training and fine-tuning the model for optimal performance
• Developed: Created user-friendly interface using Flask to display the model'spredictions and explanations.
• Deployed: Deployed the application on an Azure virtual machine.

---------------------------------------------------------------------------------------------------------------------------------------------------------

Dell TReX Managed Capacity, Client - Dell Technologies (01/2022 – 08/2022)

This recommendation engine project involved Conducting data preprocessing onlarge datasets for a recommendation engine project, optimizing ML inputs.Collaborated with stakeholders and teams to apply K-Means, DBSCAN, andhierarchical clustering for grouping. Developed a testing pipeline with A/B Testingand automation to evaluate model responses.
Key Contributions:
• Data Preprocessing Optimization: Enhanced preprocessing for large datasets to optimize ML inputs.
• Clustering Algorithm Application: Applied K-Means, DBSCAN, and hierarchical clustering for efficient user and item grouping.
• Testing Framework Development: Created a testing pipeline with A/B Testing for precise model evaluation.
• Stakeholder Collaboration: Collaborated across teams to align project goals.
• Enhanced Risk Monitoring: Improved risk monitoring in DellTechnologies' recommendation engine project.