About the role
## About the role Humyn Labs is hiring a Head of Infrastructure for its Voice AI business: the leader who owns the platform that captures, processes, evaluates and delivers speech data to frontier AI teams at scale. A builder-leader — sets the architecture, stays hands-on, and builds the team that runs it. **Location:** Bengaluru, Karnataka, India · **Employment:** full-time, on-site · **Experience:** 10+ years, 3+ leading teams ## About Humyn Labs Humyn Labs builds the intelligence layer for physical-world AI. It converts human action — sound, sight, movement and touch — into training-grade multimodal data, captured across 20+ countries in India, Southeast Asia, Latin America and the Middle East: the real environments where physical AI deploys, not the labs where it is built. Two products sit on that capture network — structured voice datasets for frontier teams building speech-to-text, speech-to-speech and multimodal models, and an egocentric video corpus shipped to robotics and world-model teams with labels such as 6-DoF head pose, 3D hand keypoints, depth and object tracks. The data is not just collected; it is evaluated, defended and production-ready. ## What you will do - Own the end-to-end infrastructure behind the voice data business: capture from contributors in 20+ countries, ingestion, storage, processing, quality checks and delivery to client AI teams. - Architect high-throughput audio pipelines — normalisation, resampling, voice-activity detection, segmentation, diarisation, ASR-assisted pre-annotation and automated quality scoring — across very large volumes of audio hours. - Run GPU inference and evaluation at scale: batch ASR and speech-model runs (Whisper and open-source models) for auto-annotation and benchmarking, with clear ownership of cost per delivered audio hour. - Build the data platform that makes every delivered hour traceable and reproducible: dataset versioning, lineage, manifests and metadata (speaker, language, dialect, consent). - Own security, privacy and compliance for speech data: consent tracking, PII handling, encryption, access control and per-client data isolation, ready for enterprise security reviews. - Set reliability and cost discipline — SLOs, observability, on-call, capacity planning and cloud spend. - Hire and lead the infrastructure team (data, platform and DevOps engineers) and set its engineering standards. - Partner with research, operations and client delivery to turn a client's dataset specification into a running pipeline quickly. ## What you need - 10+ years in infrastructure, platform or data engineering, including 3+ years leading engineering teams - Has built and run large-scale data or ML pipelines in production on AWS or GCP - Production experience with speech/audio, media processing or ML inference infrastructure (GPU workloads, model serving) - Strong with Kubernetes, infrastructure-as-code (Terraform or similar) and distributed processing (Spark, Ray, Beam or similar) - Still hands-on: designs and reviews architecture and can write production code ## Nice to have - Experience at an AI training-data, data-labelling or speech AI company. - Hands-on with speech tooling — Whisper, NeMo, Kaldi, pyannote, FFmpeg/SoX. - Data privacy and compliance work (DPDP Act, GDPR, SOC 2 / ISO 27001). - Has scaled infrastructure through a startup's growth phase. ## Working model - Full-time, on-site in Bengaluru.
Skills: Infrastructure leadership, AWS / GCP, Kubernetes, Terraform, Data pipelines, Audio processing, GPU inference, Whisper / ASR, Spark / Ray, Data security, Observability, Team building