Rohith Iyengar

About Rohith Iyengar

I am a Machine Learning Engineer driving innovation in Agentic Orchestration and…

Experience

1. Machine Learning Engineer i

   HP

   Feb 2025 — Present · Palo Alto, CA, US

   Led the development of an Agentic Orchestrator, an LLM–driven planner coordinating function-calling pipelines and real-world task execution, facilitating Agentic AI development at HP• Architected multi-turn orchestration datasets covering 30+ realistic tool-use scenarios; designed and token-optimized model outputs to improve reasoning accuracy and reduce latency by 50%+• Built and deployed end-to-end SFT, QAT, and LoRA training pipelines from scratch across Qwen 2.5/3, Phi 3.5/4, Phi Silica, and OpenAI gpt-oss-20B architectures — achieving >97% accuracy and sub-1-second response times for single-turn reasoning• Played a key role in developing a unified evaluation and benchmarking pipeline supporting Torch, ONNX, and OpenVINO backends — implemented Aphrodite’s PagedAttention optimization, resulting in up to 86% faster inference and significantly accelerating overall evaluation throughput• Implemented QAT with ONNX quantization, delivering best-performing low-latency model to date• Independently architected HP’s AWS cloud environment for distributed LLM training — including 8-GPU instances, private subnets, NAT gateways, and EC2 Connect endpoints — establishing the ML platform used by the entire team

Education

• Mission San Jose High School

  High School Diploma

• San José State University

  Bachelor of Science - BS, Computer Science