Akshat Mahajan

WORK HISTORY

EDUCATION

ABOUT AKSHAT MAHAJAN

AI Engineer at IBM with overall 4.5+ years of experience building production-grade backend systems and scalable GenAI applications. I started with backend engineering — designing APIs, handling failures, optimizing performance, and building systems that are reliable and maintainable. Over the last 2.5 years, I’ve applied that foundation to GenAI, focusing on building end-to-end AI systems that actually work in production, not just prototypes. At IBM Watsonx Client Engineering, I design and deploy RAG and agent-based systems for enterprise clients. My work spans system design, retrieval optimization, and scalable deployment using FastAPI, Docker, and cloud platforms. Key work includes: Built a multi-agent RAG system (CrewAI) for a banking client handling complex multi-hop queries using hybrid retrieval, reranking, and tool orchestrationDeveloped end-to-end RAG pipelines with Milvus, Elasticsearch, and query rewriting, improving retrieval quality and system reliability in productionFine-tuning LLMs (Phi-3-mini using QLoRA) with MLflow-based experiment tracking and evaluation (BLEU, ROUGE)Deploying scalable AI services on IBM Cloud Code Engine and AWS Core focus areas: GenAI systems · RAG architectures · Agentic AI · Backend engineering · Distributed systems · LLMOps Tech stack: Python, FastAPI, Docker, Kubernetes, REST APIs, LangChain, LangGraph, CrewAI, Milvus, Elasticsearch, Hugging Face, AWS, IBM Watsonx I’m particularly interested in roles where AI meets real backend complexity — building scalable, reliable, and production-ready intelligent systems.