Anurag Shukla

As a foundational member of the engineering team, I joined Karya (formerly a project at Microsoft Research) post-transition from Microsoft Research India to scale the technology and the team from the ground up. I lead platform strategy, architecture, and development across all systems. What I do: Team & Infrastructure Scaling: Scaling the engineering (30+ members) team while building the robust, multilingual data annotation infrastructure to support over 100,000 users globally. Platform Strategy & Development: Spearheading platform strategy, infrastructure, and feature development across web, Android, and backend systems, building the core technologies that form the foundation of Karya’s platform. PaaS Product Leadership: Leading the development of our Platform-as-a-Service (PaaS) offering. This enables partners to design and deploy custom digital labor pipelines on our modular infrastructure. Cross-Functional Leadership: Acting as interim Head of Support, building a 6-member engineering support team and its end-to-end operational processes from scratch.

Focused on translating advanced machine learning and AI research into real-world applications with significant social impact. I was a foundational contributor to projects that resulted in spin-outs and global deployments. Key Contributions: On-Device ML: Enabled fast, on-device Neural Machine Translation (NMT) on Android by integrating SentencePiece with powerful models like mT5. Created an open-source framework (Project INMT-lite, https://github.com/microsoft/INMT-lite) that can train and deploy NMT models on android Project Karya (Foundational Contributor): Core contributor to the initial platform, which later spun out as Karya. Led user research and field studies to inform platform design, improve usability, and drive worker engagement for underserved rural populations. MLOps & LLMs: Built ML pipelines for Kahani, an LLM-based storybook generator, and integrated them into Karya’s infrastructure, ensuring debugging and production readiness. Health-Tech ML: Developed and optimized ML/statistical models for SmartKC, a low-cost, portable device for the early detection of Keratoconus.

Gold Medal for Research and Innovation Gold medal for Transformational Leadership