Hello, I'm Vikram, an Applied AI Engineer with 7+ years of production engineering experience, specializing in Agentic RAG, LangGraph multi-agent systems, and LLM apps. I build end-to-end AI systems on Claude & Gemini APIs with a strong foundation in Python, FastAPI, React, and TypeScript.
About me
I'm an Applied AI Engineer with 7+ years of production engineering experience, focused on Agentic RAG, LangGraph multi-agent orchestration, and LLM application development in regulated enterprise environments.
At Scotiabank, I contributed to a production Agentic RAG platform serving 500+ internal branches. I designed agent orchestration across a LangGraph pipeline (planner, retrieval agents, tool-calling, iterative refinement), implemented hybrid retrieval (Apache Solr BM25 + vector) with MMR re-ranking, a two-level Redis caching layer, SSE streaming, and an asynchronous guardrails layer with PII scrubbing and prompt-injection interception. Retrieval quality is evaluated offline with RAGAS and LLM-as-a-judge scoring, with LangSmith tracing for end-to-end observability.
I build and evaluate end-to-end AI systems using Claude and Gemini APIs on a foundation of Python, FastAPI, React, and TypeScript. I hold the AWS Certified Developer – Associate and Anthropic Academy certifications in the Claude API and Model Context Protocol (MCP). Delivering AI in a regulated fintech context has shaped how I think about responsible AI — auditability, human escalation, and guardrails as first-class concerns.
Let's build something useful together.
My projects
Deep Research AI Agent
4-stage autonomous research pipeline (Planner → Parallel Researchers → Critic → Writer) that synthesizes structured, cited briefs in ~90s. Critic-driven rework loop with LangGraph state graphs, async parallel execution via asyncio.gather, and an LLM-as-a-judge evaluation framework scoring relevance and citation quality.
https://huggingface.co/spaces/vikraamm95/deepresearcher- Python
- LangGraph
- Gemini 2.0
- Gradio
- Tavily
- LLM-as-Judge
AI Ticket Classifier
Governed LangGraph pipeline that classifies support tickets by category, priority, and routing team with sentiment + confidence scoring. Safety layer performs PII redaction and prompt-injection detection; structured outputs validated via Pydantic. FastAPI + SQLAlchemy backend with a React 19 / TypeScript / Vite frontend and LLM-as-a-judge evaluation.
https://github.com/vikraammravi/AI-ticket-classifier- Python
- FastAPI
- LangGraph
- Claude API
- React 19
- Pydantic
My skills
- Python
- TypeScript
- JavaScript
- SQL
- Claude API
- Gemini API
- LangChain
- LangGraph
- LangSmith
- RAG
- Agentic AI
- Multi-Agent Systems
- MCP
- RAGAS
- LLM-as-a-Judge
- Hybrid Retrieval (BM25 + Vector)
- MMR Re-Ranking
- Vector Embeddings
- Prompt Engineering
- SSE Streaming
- PII Redaction
- Pydantic
- FastAPI
- Node.js
- Express.js
- SQLAlchemy
- GraphQL
- REST APIs
- React
- Next.js
- Redux
- Vite
- Apollo Client
- Tailwind CSS
- PostgreSQL
- Redis
- Apache Solr
- Docker
- Kubernetes
- AWS
- GitHub Actions
- Jenkins
- Pytest
- Jest
- React Testing Library
- Playwright
My experience
Senior Full Stack Developer | Applied AI — Scotiabank
Contributing to enterprise-scale Applied AI initiatives — designing agentic systems, retrieval pipelines, and evaluation frameworks for production use in a regulated fintech environment. Focused on responsible AI patterns including guardrails, observability, and human-in-the-loop review.
Aug 2021 - PresentMaster of Engineering — University of Windsor
Pursued a Master of Engineering, building a strong systems-thinking foundation that complements my applied AI work today.
Apr 2019 - Aug 2020Programmer — Ionixx Technologies
Drove the migration of a legacy locator system to Next.js, achieving a 30% improvement in page load times. Optimized Node.js backend performance with Redis caching, reducing query times by ~40%. Built Express.js middleware services and a real-time Project Locator UI with interactive maps.
Jun 2017 - Apr 2019Bachelor of Engineering — Anna University
Earned a Bachelor of Engineering, building a strong analytical foundation that shaped my path into software and, later, applied AI.
Jun 2013 - May 2017Certifications
AWS Certified Developer – Associate
Amazon Web Services
Building with the Claude API
Anthropic Academy · 2026
Model Context Protocol (MCP)
Anthropic Academy
Contact me
Reach me at vikraamravi@gmail.com or send a message below.