AI Engineer - Generative AI & LLMs

Vankudoth Aravind Pawar

AI Engineer

+1-203-***-**** *********************@*****.*** LinkedIn New Heaven, CT (Open to Relocate) SUMMARY

AI Engineer with nearly 4 years of experience developing and deploying Generative AI and Large Language Model applications across financial services and enterprise environments. Experienced in building Retrieval-Augmented Generation pipelines, AI assistants, multi- agent workflows, and intelligent automation solutions using Python, Hugging Face Transformers, LangChain, LangGraph, CrewAI, FastAPI, and AWS Bedrock. Skilled in LLM fine-tuning, prompt engineering, REST API integration, cloud-native AI development, and MLOps, delivering scalable, production-ready AI applications that improve operational efficiency and business decision-making. TECHNICAL SKILLS

Programming Languages: Python, SQL, Shell Scripting Generative AI & LLM Engineering: GPT-4, Claude, Gemini, Llama, Hugging Face Transformers, LangChain, LangGraph, CrewAI, LlamaIndex, Retrieval-Augmented Generation (RAG), Agentic AI, Prompt Engineering, Function Calling, Model Context Protocol (MCP), LoRA, PEFT, LLM Fine-Tuning, LLM Evaluation, Vector Embeddings Vector Databases & Knowledge Retrieval: Pinecone, FAISS, ChromaDB, Weaviate, Hybrid Search, Semantic Search, Vector Search, Semantic Retrieval, Knowledge Graphs, Document Intelligence, Reranking AI Development & MLOps: MLflow, Weights & Biases, Docker, Kubernetes, GitHub Actions, Jenkins, CI/CD, Model Deployment, Model Serving, Model Monitoring, Experiment Tracking

Cloud Platforms & Data Engineering: AWS (Bedrock, SageMaker, Lambda, Glue, S3, ECR, Redshift), Azure AI Studio, Azure OpenAI Service, Azure Machine Learning, Google Vertex AI, Apache Spark, PySpark, Apache Kafka, Apache Airflow APIs & Software Engineering: FastAPI, Flask, REST APIs, JSON, Async APIs, Microservices Databases: PostgreSQL, MySQL, SQL Server, MongoDB, Amazon RDS, Amazon Redshift Development Tools: Git, GitHub, Linux, Bash

Visualization: Power BI, Tableau, Plotly, Matplotlib, Microsoft Excel EXPERIENCE

AI Engineer JPMorgan Chase & Co, USA Jan 2026 – Present

• Developed Agentic AI applications using LangGraph, CrewAI, LangChain, AWS Bedrock, and Model Context Protocol (MCP) to automate financial compliance, enterprise search, and internal support workflows, reducing manual effort by 30%.

• Built Retrieval-Augmented Generation (RAG) pipelines using Pinecone, FAISS, semantic search, document chunking, and reranking techniques to improve enterprise knowledge retrieval accuracy by 38%.

• Developed LLM-powered financial intelligence assistants supporting regulatory interpretation, policy search, document reasoning, and enterprise knowledge discovery through secure, grounded AI responses.

• Fine-tuned Hugging Face Transformer models using LoRA and PEFT while implementing automated evaluation pipelines for hallucination detection, response quality, and factual consistency, improving answer reliability by 28%.

• Developed RESTful AI APIs using FastAPI, Docker, Kubernetes, AWS Bedrock, and Amazon SageMaker to deploy scalable LLM applications supporting secure, low-latency, real-time inference.

• Integrated MLflow, Weights & Biases, prompt evaluation, model monitoring, and experiment tracking to improve deployment reliability, model performance, and production observability.

• Collaborated with engineering, cybersecurity, legal, and product teams to integrate AI solutions into enterprise applications while supporting security, compliance, and business requirements. AI Engineer Accenture, India Jun 2021 – Jul 2024

• Developed enterprise Generative AI assistants using LangChain, GPT models, and Retrieval-Augmented Generation (RAG) to automate document understanding, enterprise search, and knowledge management.

• Implemented LLM evaluation pipelines using automated benchmarks, prompt evaluation, tracing, and human feedback loops to improve response quality, reliability, and production performance.

• Optimized prompts and evaluated LLM responses using BLEU, ROUGE, and LLM-as-a-Judge methodologies, improving response quality by 20%.

• Built intelligent document processing solutions capable of extracting, classifying, and summarizing structured and unstructured enterprise documents using Hugging Face transformer models.

• Developed scalable data pipelines using Apache Spark, PySpark, AWS Glue, Apache Airflow, and Amazon Redshift to process enterprise knowledge sources for Retrieval-Augmented Generation (RAG) and LLM applications.

• Automated AI deployment workflows using Docker, Kubernetes, GitHub Actions, MLflow, and CI/CD, improving deployment consistency and reducing release cycles.

• Integrated LLM-powered services with enterprise applications through REST APIs and microservices while collaborating with enterprise architects to implement Responsible AI practices, secure data access, and governance. PROJECTS

Enterprise Agentic AI Platform

Technologies: LangGraph, CrewAI, LangChain, GPT-4, AWS Bedrock, Pinecone, FastAPI, Docker

• Developed a multi-agent AI platform using LangGraph, CrewAI, and LangChain to automate KYC verification, AML investigations, enterprise search, compliance analysis, and IT support workflows, reducing manual effort by 40%.

• Implemented Retrieval-Augmented Generation (RAG) pipelines using Pinecone, semantic retrieval, document chunking, reranking, and contextual reasoning to improve enterprise knowledge retrieval accuracy by 40%.

• Developed secure REST APIs using FastAPI, AWS Bedrock, Docker, and Pinecone to enable scalable LLM inference, semantic search, and intelligent agent collaboration, contributing to approximately $1.5M in annual operational savings. Enterprise AI Knowledge Assistant

Technologies: AWS Bedrock, Claude, LangChain, LlamaIndex, FastAPI, Pinecone

• Developed an enterprise knowledge assistant capable of answering organization-specific questions using Retrieval-Augmented Generation (RAG), semantic search, contextual retrieval, and document reasoning across millions of enterprise records.

• Implemented prompt management, hallucination detection, LLM evaluation pipelines, conversation memory, & model monitoring to improve factual accuracy, response consistency, & production reliability while reducing support resolution time by 45%. EDUCATION

Master of Science in Artificial intelligence May 2026 University of Bridgeport, Bridgeport, USA

Contact this candidate