Ayush Kadam

Architected enterprise-grade agentic AI platform using Azure AI Agent Service, LangGraph, and CrewAI orchestrating specialized agents for IT ticket resolution, incident triage, and resolution recommendation; deployed across 12 clients autonomously handling L0/L1 tickets with 65% reduction in resolution time and 80% reduction in manual effort through coordinated multi-agent execution. Designed distributed multi-agent workflows with inter-agent communication patterns, request routing algorithms, and DAG-based execution for coordinated reasoning; integrated Azure OpenAI for LLM-powered reasoning, Azure AI Search for knowledge grounding, and enterprise systems (ServiceNow, SAP, PostgreSQL, Redis, Azure Service Bus); engineered observability tracking execution paths, decision flow, and operational metrics. Developed data integration workflows for large-scale telecom ERP migration involving 500K+ billing/financial records; implemented automated profiling, validation, and quality checks that reduced post-migration cleanup effort by ~30%.

Associate Software Engineer – Agentic AI & Platform Engineering at Capgemini (2025-07 – Present)

• Architected enterprise-grade agentic AI platform using Azure AI Agent Service, LangGraph, and CrewAI orchestrating specialized agents for IT ticket resolution, incident triage, and resolution recommendation; deployed across 12 clients autonomously handling L0/L1 tickets with 65% reduction in resolution time and 80% reduction in manual effort through coordinated multi-agent execution.
• Designed distributed multi-agent workflows with inter-agent communication patterns, request routing algorithms, and DAG-based execution for coordinated reasoning; integrated Azure OpenAI for LLM-powered reasoning, Azure AI Search for knowledge grounding, and enterprise systems (ServiceNow, SAP, PostgreSQL, Redis, Azure Service Bus); engineered observability tracking execution paths, decision flow, and operational metrics.
• Developed data integration workflows for large-scale telecom ERP migration involving 500K+ billing/financial records; implemented automated profiling, validation, and quality checks that reduced post-migration cleanup effort by ~30%.

SDE Intern – AI & Data at Samsung PRISM R&D (2024-02 – 2024-06)

• Trained ML models to generate immersive VR environments from event photo/video galleries as part of Samsung PRISM's Remini PoC — a tool enabling users to relive moments like weddings and birthdays in virtual reality.
• Built the Unity-based VR interface for the Remini app and developed Python data pipelines to preprocess and structure training datasets, reducing manual data preparation effort by ~40%.

B. Tech in Computer Science & Engineering (Business Systems) – VIT Vellore (2021-09 – 2025-05)