Vikas M.

I design and deploy production-grade AI systems, not just prototypes. If you're building with LLMs, RAG, or AI agents, the biggest challenge isn’t getting something to work. It’s making it reliable, scalable, and actually useful in the real world. That’s where I come in. Over the last 20+ years, including engineering leadership roles at Amazon and Facebook, I’ve built and led systems at scale. Today, I work with founders and teams as a fractional CTO and hands-on engineer to take AI products from idea to production. Recent work includes: • RAG systems over large document sets with fast, accurate retrieval for a large HIPPA Compliant Medical Information Product • AI agents automating workflows (research, reporting, decision support) for a Financial product • End-to-end LLM platforms with real users, memory, and structured outputs • Computer vision systems for image and video analysis for a GDPR and SOC2 Compliant Enterprise Workflow Engine Product • Full-stack AI applications (Next.js, Node.js, Python, AWS) What you can expect working with me: • Clear system design, not guesswork • Strong focus on performance, latency, and cost • Clean, maintainable architecture • Direct communication and fast iteration Typical clients: • Startups building AI products • Teams stuck after an initial prototype • Founders who need technical leadership without hiring full-time If you’re serious about building a real AI product, happy to talk. Core skills • Retrieval Augmented Generation (RAG) • Large Language Models (LLM) • AI Agents / Agentic Workflows • Prompt Engineering • LLM Orchestration Frameworks • LangChain • LlamaIndex • OpenAI API • Anthropic APis Backend • Node.js • Python • FastAPI Infra • AWS, Google Cloud, Azure • Vector Databases (Pinecone / Weaviate / FAISS) • Docker Frontend • Next.js • React PROJECTS/ PORTFOLIO **Project 1: AI Document Intelligence (RAG)** Built a production-grade RAG system over large document datasets. • Implemented document ingestion, chunking, and embedding pipeline • Designed fast retrieval with high relevance and low latency • Integrated structured outputs for downstream workflows • Stack: OpenAI, vector DB, Node.js, Python, AWS Result: Delivered a scalable system capable of handling real user queries with consistent performance. **Project 2: AI Agent Workflow System** Designed and built an AI agent system to automate multi-step workflows. • Async task execution with agent orchestration • Integrated APIs, document generation, and decision flows • Built UI + backend for real-time interaction • Stack: OpenAI, Node.js, Next.js, AWS Result: Reduced manual effort significantly and enabled automated execution of complex tasks. **Project 3: AI + Computer Vision** Developed a computer vision system for image recognition and classification. • Built training and inference pipeline • Integrated with backend services and storage • Designed for scalability and accuracy Result: Enabled automated recognition with high accuracy across varied inputs.