Abhijat Mishra — Lead ML Engineer

LLM / Gen AI Engineer | Infosys | Sep 2021 – Present I build agentic AI systems that turn legacy banking workloads into cloud-ready micro-services and conversational products. Generative AI Engineer with expertise in AI-powered legacy modernization, specializing in COBOL to Java migration and monolithic to microservices transformation. Skilled in leveraging LLMs (GPT-4, Llama2), LangChain, LlamaIndex, and vector databases (ChromaDB, Pinecone) to automate code conversion, dependency analysis, and API extraction. Developed and deployed AI-driven modernization tools on AWS (Lambda, Bedrock, EC2, S3, RDS), reducing manual effort by 50% and improving transformation speed. Passionate about integrating AI, cloud computing, and automation to drive efficiency in enterprise modernization. 🏦 Bank Helpdesk GenAI (2025) Built a conversational RAG assistant: ingested 50 k SOP PDFs into a vector store, wired GPT-4o function-calling to core banking APIs for balance look-ups, card blocks and dispute status. First-contact-resolution jumped 62 → 88 %. Added JSON-schema guardrails ⇒ 0 prod incidents; Streamlit trace-viewer shows every retrieval & tool call. 🔧 GenAI Utilities Metadata-scanner agent draws inter-service dependency graphs—saving $10 k/yr in manual analysis. Event-trigger generator converts legacy batch configs into AWS Step Functions, shaving two sprints per migration. ☁ MLOps Dockerised pipelines, served via FastAPI + AWS Lambda, observed with OpenTelemetry; p95 latency 250 ms at 100 RPS. Stack: GPT-4o, Llama-2/3, LangChain, LangGraph, FAISS, OpenSearch, Python, FastAPI, Spring Boot, Docker, AWS (Lambda | EC2 | S3 | Step Functions), Streamlit, GitHub Actions.

Stackforce AI infers this person is a Fintech-focused AI Engineer with expertise in cloud and legacy system modernization.

Location: Kanpur, Uttar Pradesh, India

Experience: 4 yrs 6 mos