AI/ML Tech Lead specializing in production-grade Generative AI, multi-agent orchestration, and cloud-native distributed systems.
I design and build secure, governed, and enterprise-ready AI systems that scale. My expertise spans building low-latency LLM & RAG pipelines, custom agent gateways (using MCP and ADK), and automated MLOps pipelines across AWS and GCP.
- 🧠 Enterprise AI Engineering: Scalable RAG, LLM orchestration, and production-grade multimodal systems.
- 🤖 Agentic Workflows: Multi-agent system design, Model Context Protocol (MCP) implementations, and ADK tool-use.
- ☁️ Cloud Architecture & MLOps: Cost-efficient GPU/CPU batch inference, IaC (Terraform/CDK), and automated CI/CD retraining.
- 🧬 Healthcare AI: Deep learning models for medical imaging, radiology pipeline orchestration, and clinical deployments.
- 🔐 Governance & Observability: Secure agent control planes, RBAC policy enforcement, and transaction telemetry.
| Area | Technologies |
|---|---|
| Agentic AI | MCP ADK ReAct LangGraph LangChain Multi-Agent Orchestration Tool Routing |
| AI / Deep Learning | LLMs (Claude/GPT/Llama/Gemma) PyTorch TensorFlow HuggingFace U-Net DenseNet |
| MLOps & Infrastructure | AWS (SageMaker, Bedrock, Lambda, S3, ECS) GCP (Vertex AI, Cloud Run) Terraform CDK |
| Backend & Platforms | Python FastAPI Flask Docker Kubernetes Microservices GitHub Actions |
Collaborative AI Agents for Secure Data Analytics
Built a collaborative multi-agent platform for enterprise data intelligence, orchestrating specialized agents (SQL Planner, Executioner, Data Critic, and Visualizer) to securely query, validate, and analyze multi-source database systems with zero SQL-injection risk.
Secure Control Plane for Enterprise Agents
Designed a secure control plane for agent-tool interactions, implementing role-based access control (RBAC), security policy enforcement, and observable agent transaction logging across MCP clients and servers.
AWS GPU Batch Inference
Engineered a medical image processing pipeline using AWS Batch and GPU orchestration (MRIQC + FastSurfer) to automate radiology workflows, reducing processing latency by ~60%.
Multi-Agent RAG & Policy Enforcement
Built a secure multi-agent reasoning network for automated threat intelligence and compliance analysis. Combines secure document ingestion with strict verification control loops.
Deep Learning for Clinical Decision Support
- IVF Embryo Grading: Developed a multi-head DenseNet model for structured embryo scoring.
- Lung Opacity Segmentation: Built an attention-based U-Net model for automated pneumonia detection on CXRs.
- 🏆 Top 2% — Kaggle Brain Tumor Radiogenomic Classification Challenge
- 📄 IEEE Publication — Electrical Field Distortion Around Man-Made Objects
🚀 Building scalable, reliable, and governed AI systems for real-world impact.