Akash Chaute

GenAI Engineer with 2+ years of production experience building agentic AI systems at Comcast (via Accenture), serving 4,000+ users. Expert in multi-agent orchestration (LangGraph, LangChain), RAG pipelines, and scalable LLM systems. Delivered AI automation cutting 70–80% manual engineering effort.

Claude Certified

Architect (Anthropic) · AWS Certified · Experienced in AI product development and LLM evaluation · Top 1.7% in JEE 2020 out of 1.2M candidates.

AI/ML Computational Science Engineer Analyst at Accenture (2024-06 – Present)

Genlite-Genie — Enterprise GenAI Platform | Comcast · 4,000+ Users

• Accelerated defect resolution by 70% across 4 teams by architecting a production multi-agent LangGraph orchestration framework with 4 specialized agents (GitHub, Jira, AST, Fix Generator).
• Automated the full triage → root-cause → PR generation workflow, eliminating manual triage for ~400 engineers.
• Cut PR generation latency by 40% via ThreadPoolExecutor-based concurrent AST analysis and GPT-4o-powered code generation and fix synthesis.
• Eliminated 3–4 hrs of manual reporting per sprint for 5 Scrum teams (~500 users) by building an agentic LangGraph + FastAPI pipeline with cron-triggered email collection and reply aggregation.
• Auto-generated executive PowerPoint decks containing sprint health scores, velocity trends, and blocker analysis, slashing reporting overhead by 80% per sprint cycle.
• Delivered automated sprint intelligence, saving ~500 users from repetitive reporting each sprint.
• Restored 100% uptime for a production AI code-generation pipeline serving 3,000+ users by resolving a critical Azure Service Bus bottleneck.
• Decoupled a monolithic queue into independent async image and JSON consumer services.
• Eliminated throughput failures under high concurrent load, improving overall pipeline throughput by 3×.

Research Intern → Research Member at IvLabs – VNIT, Nagpur (2022-06 – 2023-10)

• Developed an autonomous obstacle avoidance pipeline for the Sahayak medical robot using ROS, A*, RRT, and B-spline trajectory optimization, achieving ~10–15% shorter and smoother planned trajectories.
• Built and evaluated 20+ custom ROS simulation environments, executing 15+ navigation experiments with static and moving obstacles to assess obstacle avoidance performance.
• Optimized SLAM loop closure by benchmarking 4 place recognition models.

B.Tech – National Institute of Technology, Nagpur (2020-12 – 2024-05)