**2024**Managing a team focused on synthetic data generation, LLM evaluation, and experimental data annotation techniques.**2023**Managed a fullstack SWE & MLE team working on Coral, Cohere's flagship enterprise knowledge assistant. Led efforts that span from backend, infra & data, all the way to models, prompts, and UI. Also kept an emphasis on building a diverse, safe, fun team that supports engineers in delivering across the stack and across SWE-MLE boundaries.**2022**Led & managed a team of machine learning engineers, working on exploratory applications of large language models (LLMs). Areas of work included conversational AI, learning from human feedback, and efficient evaluation of LLMs. Built a mixture of open source libraries, generative AI models, API endpoints, and human-LLM interactive data generation pipelines. Also focused on championing async-first best practices and diverse & inclusive communities.

Part of Vector Institute's affiliate program for researchers in AI & related domains of application.

Adjunct Professor in the Dept. of Materials Science and Engineering. Currently focused on the intersection of large language models and materials informatics.

Led and managed a team of scientists towards creating machine learning products serving millions of users across the globe.Owned a portfolio of ML initiatives spanning across the domains of machine vision, natural language processing, MLOps, and responsible AI. Relevant technologies used include Tensorflow, MLflow, and many services within AWS. Co-chair of the Asian and Pacific Islander Employee Resource Group (ERG) at Xero, growing its membership by 6x to over 100 employees in its first year. Also a member of Xero's Canada Leadership Team, a small group of senior leaders & executives who coordinate country-wide initiatives.

Used machine learning to solve industry challenges in materials and chemicals development as a customer-facing team lead. Championed key internal R&D efforts, pioneered the improvement of data science workflows, and co-authored multiple academic communications (e.g., conference tutorials, journal articles).Developed, prototyped, and delivered a variety of new technologies using Python, Jupyter, Keras, Git, JIRA, AWS, and related tools. Co-invented two US patents, delivered AI workshops for hundreds of industry and academic professionals, led technical strategy for multi-year seven-figure contracts, and mentored numerous data scientists and engineers.

Researched computational predictive synthesis techniques for inorganic materials via machine learning / natural language processing / data mining methods as a member of the Olivetti group in the Dept. of Materials Science and Engineering. Developed a scalable, automated machine learning pipeline from the ground-up, using modern deep learning libraries (e.g., Keras, Tensorflow), standard software engineering components (e.g., Git, MongoDB, RabbitMQ, CircleCI), and modern project management workflows (e.g., Asana, Slack).Academic contributions include: domain-specific word embeddings and named entity recognition models, machine-learned insights into nanomaterial synthesis outcomes, and representation learning of materials synthesis routes using deep generative models.

Designed and implemented a variety of machine learning approaches to solve large-scale business problems. Techniques used include recurrent/convolutional neural networks, unsupervised word embeddings, transfer learning, and domain-specific named entity recognition.Also wrote full-stack code (Python) for machine learning microservices (AWS, Flask, Docker), provided data visualizations (Plotly, Jupyter, D3), and worked with a team using agile development practices (JIRA).