Hi, I am
Bogam Abhishek
Building next-generation Agentic AI systems — from multi-agent RAG pipelines to enterprise LLM orchestration — that solve real-world problems at scale. Passionate about making AI work smarter, not harder.
The Engineer Behind the AI
Hey! I'm Bogam Abhishek, an Associate Software Engineer at Tech Mahindra, where I work as a Gen AI Engineer building next-generation agentic AI systems that drive real enterprise impact.
My focus is on LLM orchestration, Agentic RAG architectures, and multi-agent systems — intelligent pipelines that don't just retrieve information, but reason, plan, and act on it.
I've delivered systems reducing manual effort by 45%, cutting API costs by 25%, and improving query accuracy by 30% — all in production telecom environments.
- 📍 Andhra Pradesh, India
- 💼 Associate Software Engineer @ Tech Mahindra
- 🎓 B.Tech CS — MITS (GPA: 8.42/10)
- 📧 bogamabhishek27@gmail.com
Agentic AI
Multi-agent RAG with planning, reasoning, and sub-query decomposition.
LLM Infra
On-prem & hybrid deployment using vLLM, OpenVINO, Groq and HuggingFace.
Award Winning
3rd place @ Microsoft × TechM Ideathon. Awarded @ Microsoft Agentic AI Hackathon.
Cloud Native
Azure-certified. Deployed production AI systems on Azure in rapid 4-day cycles.
Where I've Made Impact
Production-grade AI systems for real enterprise challenges.
- → Developed and deployed an Agentic RAG system for automated 4G/5G (3GPP) test case generation, reducing manual effort by 45%.
- → Designed multi-step reasoning workflows using planning-based task decomposition, improving complex telecom query accuracy by 30%.
- → Implemented domain-specific embedding segmentation (AMF, SMF, Core Network) with intelligent retrieval routing, reducing latency by 35% and improving semantic relevance.
- → Engineered a scalable Flask backend integrating on-prem LLM inference (vLLM, OpenVINO) and hybrid API models (Hugging Face, Groq), enabling multi-model orchestration and concurrent inference.
- → Optimized modular RAG pipelines for enterprise deployment, reducing token consumption and API cost by 25%.
Tools of the Trade
Technologies I use to build intelligent, production-grade AI systems.
Things I've Built
Production-ready AI systems tackling real-world enterprise challenges.
A source-aware, memory-enabled RAG system built with FastAPI, FAISS, and LLaMA-3 via Groq for grounded responses on large technical documents. Retrieval-first architecture with domain-aware controls to enhance accuracy and minimize hallucinations.
Production-ready AI system for automated KYC validation and financial risk assessment using Flask, Azure Blob Storage, and Azure AI Document Intelligence. Integrated hybrid vector retrieval with Azure AI Search and Azure OpenAI to generate structured PDF credibility reports — deployed on Azure in just 4 days.
Multi-agent RAG system with clustered topic-based embedding generation, Sub-Query Decomposition Agent, and intelligent retrieval routing for improved semantic precision and multi-hop reasoning across complex, distributed knowledge bases.
Recognition & Achievements
Honored for technical excellence and innovative AI delivery.
Top Performance Rating — Two Consecutive Years
Received Top Performance rating for two consecutive years at Tech Mahindra for outstanding technical delivery in Gen AI engineering.
3rd Place — Microsoft × Tech Mahindra Ideathon
Secured 3rd place among 80+ teams for presenting an Agentic RAG model demonstrating real-world enterprise applicability.
Microsoft Agentic AI Hackathon Award
Awarded at the Microsoft Agentic AI Hackathon for developing a multi-agent solution using the AutoGen framework.
Pat on Back Award — Tech Mahindra
Recognized for exceptional contributions and going above and beyond the scope of assigned technical deliverables.
Certified Expertise
Industry-recognized credentials across cloud, AI, and development.
Academic Foundation
Andhra Pradesh, India
Andhra Pradesh, India
Andhra Pradesh, India
Let's Build Something Great
I'm open to exciting Gen AI opportunities, collaborations, and conversations about agentic systems, LLM architectures, or anything at the frontier of AI. Don't hesitate to reach out — I'd love to connect.
Send a Message
I typically respond within 24 hours.