Samim Kumar Patel
Rourkela
Summary
Over 5 years of expertise in Generative AI and Agentic systems, with a proven track record of leading engineering teams through complex technical deliveries. Expert in architecting Multi-Agent systems using Google ADK and Vertex AI, and implementing GraphRAG systems with Neo4j to drive high-precision enterprise solutions. Extensive Experience managing secure, scalable cloud infrastructure across GCP and AWS, backed by Professional certifications in DevOps, Security, and Kubernetes.Skilled in guiding cross-functional teams of up to 5 members, focusing on mentoring senior engineers and fostering a collaborative, high-performance culture.
Overview
7
7
years of professional experience
1
1
Certification
Work History
Agentic AI & GenAI Engineer
Samim AI Consulting
01.2023 - Current
- Engineered a sophisticated agentic multi-agent system using Google Agent Development Kit (ADK) and Vertex AI for autonomous task orchestration and dynamic decision-making.
- Implemented persistent memory management and contextual state handling to enable long-term conversation recall and personalized interactions across agents.
- Integrated Retrieval-Augmented Generation (RAG) with Vertex AI for enhanced response accuracy and grounding in enterprise knowledge bases.
- Deployed the application on Google Cloud Run with Docker containerization and Kubernetes for auto-scaling, high availability, and fault tolerance.
- Secured API endpoints using OAuth 2.0, Workload Identity Federation, and Cloud IAP (Identity-Aware Proxy) for robust authentication and zero-trust access control.
- In-House Project Team: 3
- Technologies: Google ADK, Vertex AI, Gemini, LangChain, PostgreSQL, Docker, Kubernetes, Cloud Run
System Architect
Samim AI Consulting
01.2022 - 12.2022
- Designed and developed a GraphRAG-based system using Neo4j knowledge graph integrated with LangChain to deliver more accurate, contextual, and explainable responses compared to traditional vector RAG.
- Built hybrid retrieval pipelines combining graph traversals (Cypher queries for entity-relationship reasoning) with semantic search for superior query understanding and reduced hallucinations.
- Utilized LangGraph for orchestrating complex multi-step reasoning flows and dynamic tool invocation across the agentic workflow.
- Stored interaction logs and graph updates in MongoDB for analytics, performance monitoring, and continuous improvement.
- Achieved higher output precision and structured reasoning for enterprise use cases involving interconnected domain knowledge.
- In-House Project Team: 4
- Technologies: LangChain, LangGraph, Neo4j, OpenAI/Gemini, RAG, MongoDB
Machine Learning Engineer
Samim AI Consulting
01.2020 - 12.2021
- Fine-tuned Llama 3.1 8B Instruct model on Hugging Face AutoTrain for a specialized domain-specific chatbot.
- Developed custom Python scripts to upload and manage training datasets on the Hugging Face Hub.
- Deployed the fine-tuned models on AWS EC2 instances provisioned and managed via Terraform Infrastructure as Code.
- Optimized training and inference performance through hyper parameter tuning, mixed precision techniques, and efficient resource utilization.
- In-House Project Team: 5
- Technologies: Hugging Face, Transformers, AWS EC2, Terraform
AWS System Architect
Samim AI Consulting
01.2019 - 12.2020
- Designed and implemented a complete end-to-end AWS-native CI/CD pipeline using AWS CodePipeline and AWS CodeBuild for the GraphRAG AI application.
- Automated the full software delivery process including source code checkout from Git, Docker image build, unit/integration testing, vulnerability scanning, and secure image push to Amazon ECR.
- Configured multi-stage pipeline with separate Build and Deploy stages for reliable and auditable releases across development, staging, and production environments.
- Provisioned and managed Amazon EKS (Elastic Kubernetes Service) clusters, IAM roles and supporting infrastructure entirely as code using Terraform.
- Enabled zero-downtime deployments to Kubernetes using Helm charts with rolling updates and Horizontal Pod Autoscaler (HPA) for automatic scaling based on load.
- In-House Project Team: 4
- Technologies: AWS CodePipeline, AWS CodeBuild, Amazon ECR, Amazon EKS, Terraform, Docker, Kubernetes, Git.
Education
Bachelor of Computer Science -
UTKAL University, KITS
Bhubaneswar, Odisha04-2001
Skills
- PyTorch framework
- Model evaluation
- Python programming
- TensorFlow framework
- Scikit-learn library
- Feature engineering
- Neural networks
- Machine learning
- Natural language processing
- Reinforcement learning
- Keras library
- Deep learning
- Excellent communication
- Critical thinking
- Team collaboration
Certification
- AWS Certified Solutions Architect – Associate 2022
- Google Cloud Professional DevOps Engineer 2022
- Google Cloud Professional Security Engineer 2022
- Google Cloud Associate Cloud Engineer 2021
- Certified Kubernetes Administrator (CKA) 2024
- Docker Certified Associate 2023
- Credly Digital Badges View on Credly
- LinkedIn Profile www.linkedin.com/in/samim-kumar-patel
Declaration
I, Samim Kumar Patel, affirm that all details provided are accurate and true to the best of my knowledge.
Timeline
Agentic AI & GenAI Engineer
Samim AI Consulting
01.2023 - Current
System Architect
Samim AI Consulting
01.2022 - 12.2022
Machine Learning Engineer
Samim AI Consulting
01.2020 - 12.2021
AWS System Architect
Samim AI Consulting
01.2019 - 12.2020
Bachelor of Computer Science -
UTKAL University, KITS