AI Engineer - RAG, Agentic LLMs, AWS Deployments

Ankit Patil

**************@*****.*** 646-***-**** LinkedIn Github New York City, NY (Ready to Relocate) Professional Summary

AI Engineer with 2+ years, building production RAG pipelines, Agentic LLM workflows, and Governed AI Systems on AWS. Experienced shipping Multi-tenant FastAPI microservices with LangGraph, pgvector, and OpenAI APIs. Strong background in Python, LangGraph, pgvector, Neo4j, Prompt engineering, Vector embeddings, and end-to-end Cloud deployment on AWS ECS/Fargate through CI/CD pipelines Work Experience

Invisia AI Remote, USA

AI Engineer Oct 2025 - Present

• Architected a multi-tenant RAG Knowledge Vault by using pgvector & Python ingesting 50MB+ financial files with tenant-isolated vector ranking and zero cross-tenant retrieval bleed

• Accelerated investor reporting cycles by 40%, resulting in faster compliance approvals, by orchestrating a human-in-the-loop agentic pipeline using LangGraph and GPT-4o

• Engineered ToneGuard, a citation-based safety guardrail and faithfulness suppression layer writing cryptographically immutable audit logs, reducing compliance review time

• Secured the agentic layer’s access to internal tools, resulting in fully governed AI operations, by deploying a custom MCP server and RBAC-enforced microservices using AWS ECS Fargate, Docker & GitHub Actions G5InfoTech Remote, USA

AI Intern Jan 2025 - Oct 2025

• Designed a multi-turn RAG chatbot with LangGraph and pgvector using hybrid dense/sparse retrieval and cross-encoder reranking, achieving 0.87 faithfulness on RAGAS and automating 45% of internal queries

• Reduced domain-specific hallucinations 3x by integrating Ollama local models with structured prompting, persistent memory, and exception-aware retry logic

• Implemented a Graph RAG pipeline over internal knowledge assets using Neo4j and LLM-guided traversal, improving multi-hop answer accuracy by 35% versus flat vector retrieval SafeSpace Remote, USA

Software Engineering Intern Aug 2024 – Dec 2024

• Developed FastAPI backend features and automated test suites for a production application serving early-stage users under iterative code-review cycles, resulting in a 20% drop in post-deployment bugs

• Orchestrated REST and SSE APIs with a Next.js frontend to stream real-time system states, cutting simulated data triage time from hours to minutes

Education

State University of New York, Binghamton Binghamton, NY, USA Master of Science, Computer Science Jan 2023 - May 2025 Savitribai Phule Pune University Pune, India

Bachelor of Engineering, Computer Engineering Aug 2018 - May 2022 Projects

Graph RAG Knowledge Assistant LangGraph, Neo4j, OpenAI, FastAPI, Langfuse

• Constructed entity-extraction and graph-ingestion pipelines over financial filings using Neo4j and LLM-guided traversal, improving multi-hop recall by 38% versus dense-only retrieval

• Instrumented full pipeline observability with Langfuse tracing and a RAGAS evaluation harness comparing fixed, semantic, and late-chunking strategies across faithfulness, context recall, and answer relevance dimensions Incident Zero - Multimodal Agentic Security Platform FastAPI, Next.js, Mistral, MCP, OCR, SSE

• Spearheaded a multimodal agentic pipeline using FastAPI and Mistral to automate OCR-based evidence extraction, attack-graph generation, and root-cause analysis

• Configured a structured MCP tool registry utilizing LLM function-calling and strict JSON structured outputs to guarantee predictable, type-safe agent tool execution WanderGenie AI Travel Assistant/Chatbot LangGraph, AWS Bedrock, GPT-4o-mini, pgvector, Neo4j

• Devised a multi-agent AI travel planner using AWS Nova Pro, Bedrock, GPT-4o-mini with LangGraph orchestration for coordinating planner, researcher, and packager agents for multi-step itinerary generation

• Optimized a hybrid information retrieval pipeline with RAG including chunking strategies, vector store indexing via Supabase pgvector, Neo4j & evaluated offline against relevance metrics, boosting route coherence by 40% Skills

Programming: Python, SQL, TypeScript, JavaScript

AI & LLMs: RAG, Graph RAG, Hybrid Retrieval, Agentic Systems, LangGraph, LangChain, LlamaIndex, OpenAI GPT-4o, Claude, Mistral, Ollama, Prompt Engineering, RAGAS, LLMOps, MCP Server, Cross-Encoder Reranking Databases & Vector Stores: pgvector, Neo4j, Pinecone, FAISS, ChromaDB, PostgreSQL, MongoDB Cloud & Infra: AWS ECS Fargate, Lambda, S3, EC2, SageMaker, Azure, Docker, Kubernetes, Terraform, CI/CD Frameworks: FastAPI, Flask, REST APIs, React.js, Microservices Certifications: NVIDIA Generative AI - Associate & Professional NVIDIA Agentic AI - Professional

Contact this candidate