SHIVAM
THAKUR
AI Engineer. I build AI products that actually ship.
Background
I'm an AI Engineer based in Mumbai, currently working as an AI Research Assistant at IIT Bombay. I build real AI products — not demos.
I've shipped three live products (MODA AI, Complicenter, Adally), built a community of 350+ AI professionals across India, and worked with APIs like Claude, Gemini, Replicate, and Vertex AI. I care about products that solve actual problems.
Capabilities
LLM Engineering
Expertise in RAG, prompt optimization, and fine-tuning for production-grade performance.
Multi-Agent Systems
Designing complex workflows where multiple specialized AI agents collaborate on enterprise tasks.
AI Product Dev
Full-stack implementation from ideation to live deployment using modern stacks.
Automation
Orchestrating business processes with N8N, LangChain, and custom integration layers.
Infra & Deploy
Scalable deployment on Vertex AI, AWS, and edge providers for low-latency AI apps.
Open to custom AI consulting
Live Products
MODA AI
AI fashion mockup generator converting design images into photorealistic e-commerce mockups — reducing visualization prep time by ~70%.
Complicenter
LLM compliance agent evaluating marketing creatives against advertising platform policies — reducing manual review effort by ~50%.
Adally
AI system generating YouTube content strategies, ad scripts, and campaign frameworks with consistent output across multiple marketing formats.
AI India Network
Building the largest community of 350+ AI professionals across India. Jobs board, startup directory, events hub & more.
Join CommunityYoga Pose Detector
Real-time yoga pose classification using computer vision and pose estimation. Detects and evaluates body posture live from webcam input with instant feedback.
Hand Gesture Racing
Browser-based racing car game controlled entirely by hand gestures via webcam. Real-time hand tracking maps finger positions to steering and acceleration — no keyboard needed.
Chronicle
AI Research Assistant
IIT Bombay
Designing multi-agent architectures for research automation. Built knowledge retrieval pipelines using embeddings and semantic search. Reduced manual research effort by ~40% through automated document parsing and structured information extraction workflows.
AI Intern
Aicade.io
Built an AI-powered coding tool using Claude, RAG pipelines, and tool-calling workflows. Implemented local LLM fallback using Ollama, reducing model failure rate by ~25%. Developed image generation pipelines using Mistral and Vertex AI Imagen.
ML Intern
Techluminix Pvt Ltd
Built and deployed LLM-powered chatbot systems for client websites and CRM workflows. Delivered end-to-end chatbot deployments for multiple clients, improving response accuracy by ~20% through iterative prompt refinement.
Transmission
Ready to ship something intelligent?
Open to freelance AI projects and full-time roles.
shivamthakur6888@gmail.com