Banriskhem K. Khonglah

WORK HISTORY

EDUCATION

ABOUT BANRISKHEM K. KHONGLAH

I am a Senior AI Engineer at Supermodular.ai, where I develop AI products with a current focus on Agentic AI technologies.My day-to-day work centers on building and shipping agentic systems, designing stateful, multi-step agent graphs that orchestrate tool use, reasoning, and decision-making. I leverage browser agents to automate complex web-based tasks such as information extraction, and end-to-end workflow execution. I also build evaluation frameworks to measure agent reliability, behavioral correctness and implement observability infrastructure for tracing and monitoring agent execution in production.Beyond agentic AI, I bring extensive experience in fine-tuning LLMs, prompt engineering and retrieval-augmented generation (RAG) pipelines.My background is rooted in Machine Learning research for speech and language technologies. Over the years, I have designed and developed models for Automatic Speech Recognition, Keyword Spotting, Intent-Based Natural Language Understanding, Machine Translation, and Text-to-Speech Synthesis. I have a strong understanding of sequence-to-sequence architectures, including Transformers, BERT, and GPT.Prior to my current role, I served as a Staff Engineer at NIO in Germany, where I improved the end-to-end experience of NOMI, NIO\'s in-car voice assistant, focusing on ASR applications for the automotive domain. Before that, I worked as a Senior Research Scientist at Cerence in Germany, specializing in language modeling for automotive ASR systems.I began my industry career following a postdoctoral appointment at the Idiap Research Institute in Switzerland, where I contributed to the IARPA-funded MATERIAL project (\"MAchine Translation for English Retrieval of Information in Any Language\"), conducting research in automatic speech recognition for low-resource and multilingual settings.Technologies- Automatic Speech Recognition (ASR)- Generative AI (GenAI), Large Language Models (LLM) and Fine-tuning of LLMs- Natural Language Processing (NLP)- Natural Language Understanding (NLU)- ASR + NLU end to end- Text to Speech Synthesis (TTS)- Audio Classification - Speech Enhancement (ECNR)- Speaker Recognition/ Speaker Verification/ Speaker Diarization / Multi-Speaker Separation Skills- Python, Matlab, C- PyTorch, Tensorflow- Git, Bash, AWS Sagemaker, Microsoft Azure ASR - Kaldi, HTK, Sphinx, VOCON- Hugging Face Transformers, Pipelines- LLM APIs- Prompt EngineeringGoogle & hl=en