RAKESH SAI DHEERAJ M

Project: Enterprise Knowledge Assistant (GenAI)

Domain: Enterprise Knowledge Management

• Developed a Generative AI–based knowledge assistant to enable users to query large volumes of internal documents and obtain context-aware, accurate responses.
• Implemented Retrieval-Augmented Generation (RAG) pipelines to fetch relevant information from document repositories and enhance LLM responses with grounded context.
• Worked on document ingestion and preprocessing pipelines supporting multiple formats such as PDF, DOC, PPT, TXT, and CSV.
• Designed and managed LLM workflows using LangChain and LangGraph, enabling structured reasoning, tool calling, and controlled response generation.
• Applied Prompt Engineering techniques to improve answer accuracy, reduce hallucinations, and ensure consistent response formatting.
• Built and exposed Flask-based REST APIs for query handling, document retrieval, and response generation.
• Integrated automation workflows using n8n to orchestrate document updates, indexing, and background processing tasks.

Collaborated with stakeholders to align the assistant’s responses with business-specific terminology and knowledge requirements.

Project: AI-powered Question Generation Platform

Domain: Learning & Assessment Platform

• Worked on developing a web application–based AI platform that accepts documents in multiple formats such as PDF, CSV, PPT, DOC, TXT, and other structured and unstructured inputs.
• Designed and implemented Generative AI–based pipelines to automatically generate Multiple Choice Questions (MCQs) from uploaded documents.
• Built logic to generate questions at multiple difficulty levels (Easy, Medium, Advanced) based on content complexity and learning objectives.
• Integrated LLM-driven workflows with document parsing, content chunking, and contextual understanding for accurate question generation.
• Developed and exposed Flask-based REST APIs to handle document ingestion, processing, and question generation.
• Applied Prompt Engineering and Retrieval-Augmented Generation (RAG) techniques to improve relevance, accuracy, and contextual consistency of generated questions.
• Collaborated with frontend and product teams to ensure seamless integration of AI services into the web application.

Project: Research on HR management and perform predictive models in HR sector

Client: Internal project (POC)

Domain: HR Solutions

• Performed the EDA and the distribution plots for understanding the data distribution of the given features in the given input data.
• Derived the new features by using feature engineering techniques and also performed the feature transformation.
• A end to end classification model was developed.

Project: Server Blocking Prediction

Domain: RETAIL Industry

• Developed deep learning models to predict the server blocking or not in next 25 minutes based on the features given in the data.
• Build python scripts using developed models and delivered to the client to run on their servers.
• Worked on the EDA part to understand the data distribution of the features and identified the features for the model having separable distribution of the output classes.
• As a part of EDA, build several plots using Plotly and cufflinks to demonstrate the feature distribution of feature to the client.
• Performed various feature engineering techniques and derived few features from the given data which having more weightage in the model.

Domain: Medicine Manufacture Company

• Performed the EDA and the distribution plots for understanding the data distribution of the given features in the given input data.
• Derived the new features by using feature engineering techniques and also performed the feature transformation.
• A End to End regression model was developed and delivered to the client.
• A Flask api was created for the developed model.