Artificial Intelligence Developer - RAG & LLMs

Akiko Sherman infotech Pvt Ltd, Delhi

Odisoft Technology Pvt Ltd, Odisha

SANJAY KUMAR SETHI

Artificial Intelligence Developer

EDUCATION

B. TECH (MECH)

BPUT, ODISHA

****-****

SKILLS

LANGUAGE : PYTHON.

FRAMEWORK : FLASK, DJANGO, FastAPI.

FRONTEND : HTML, CSS, BOOTSTRAP.

DATABASE : MySQL, POSTGRESQL.

VISUALIZATION : MATPLOTLIB, SEABORN.

LIBRARIES : NUMPY, PANDAS, MATPLOTLIB,REGEX,

SCIKIT-LEARN, TENSORFLOW, PYTORCH,

KERAS.

CLOUD : AWS, GCP, AZURE.

AGENTIC AI : LANGCHAIN, LANGGRAPH,CREWAI,

AGNO, AUTOGEN, ADK, N8N.

ArtificialIntelligence, MachineLearning, DeepLearning, LLMs, GenerativeAI, NLP, HuggingFace, OpenAI, RAG, Finetune, Chunking, Tokenization, Embedding, Transformer, MCP, ChromaDB, Pinecone, PromptEngineering, Git,

Jenkins, CI/CD, Docker.

707-***-****

**************@*****.***

Bangalore, India

EXPERIENCE

Artificial Intelligence Developer

DEC-2021 – MAY-2025

Artificial Intelligence Developer

JUNE-2025 – OCT-2025

• Design, Develop and Optimize conversational ai

models to Solve complex real-world problems across various domains.

• Uses programing language like python and frameworks to build, test and refines models while ensuring code quality and scalability.

• Integrates AI models into applications, deploys scalable solutions on cloud platforms, and ensures ethical, reliable, and maintainable system performance.

• Work closely with it teams to ensure seamless integration of AI solutions within existing infrastructure, oversee the deployment process and address any technical chalanges that arise.

• Stays updated with latest AI advancements, researches new techniques, and applies innovative methods to improve existing systems and develop new solutions.

Translation Automation

ChatBot

Insurance Inquiry Text Classifier

PROJECTS

• Developing and Designing a production-grade, real-time RAG-based conversational AI system for intelligent question answering over enterprise data.

• Built an end-to-end RAG pipeline using LangChain for document ingestion, cleaning, chunking, and embedding storing on vector database.

• Generated embeddings using OpenAI and stored vectors in pinecone vector database.

• Applied Top-K retrieval followed by Maximal Marginal Relevance (MMR) to optimize contextual relevance and minimize redundancy.

• Implemented Redis for session-based conversation memory.

• Exposed APIs using FastAPI, Docker and deployed on AWS EC2.

• Developing an end-to-end automation workflow using Airtable and Zapier, eliminating manual translation processes.

• Configured trigger-based data pipelines to process and update records automatically in real time.

• Implemented workflow resilience with retries, error handling, and monitoring for failed executions.

• Streamlined data synchronization and updates, significantly reducing turnaround time and operational overhead.

• Improved translation consistency and quality by standardizing inputs and outputs across the pipeline.

• Developing an end-to-end Insurance Inquiry Text Classification system to automate query categorization.

• Fine-tuned models on domain-specific insurance datasets to improve classification accuracy and contextual understanding.

• Implemented RAG (Retrieval-Augmented Generation) techniques to enhance classification by incorporating relevant contextual information from knowledge sources.

• Built and optimized text preprocessing pipelines using regex, tokenization, and normalization.

• Developed a scalable pipeline for efficient routing of user inquiries to reducing manual effort.

Contact this candidate