Leadership¶
I work where regulated, production-grade AI meets applied research. My focus is building AI teams and systems that add lasting business value instead of technical debt.
How I Lead¶
1. Business Outcomes and Reliable Production¶
At Teladoc Health, I lead machine learning teams working in a healthcare setting with high standards for trust, privacy, and compliance. In practice that means:
- Outcomes first: R&D ideas earn engineering time by showing clear product, clinical, or cost impact.
- Reliable production: End-to-end MLOps with monitoring, drift detection, privacy safeguards, and audit trails.
- Cross-functional alignment: Working closely with Product, Clinical, Legal, and Infrastructure leaders so the AI roadmap supports company goals.
2. Cost-Aware AI Architecture¶
Inference and training costs can quietly erode the value of an AI product. I design systems with cost in mind from the start:
- Right-sized models: Where the task allows, specialized self-supervised models (such as JEPA variants) can match or beat large general-purpose models at a fraction of the compute cost. Our pico-JEPA v2 work is one example.
- Everyday hardware: Memory-efficient training and fast data pipelines that run on consumer GPUs with 6–8 GB of memory, or on edge devices.
3. Connecting Research, Teaching, and Industry¶
As a Professor at UTN FRBB and leader of the BHI-Research group, I keep a direct line between new research and real-world engineering:
- Early evaluation of new ideas: We test emerging methods (Joint-Embedding Predictive Architectures, audio-video models, efficient world models) hands-on, well before they become standard tools.
- Mentoring: Supervising undergraduate and graduate researchers and helping them grow into strong ML engineers and scientists.
Core Competencies¶
| Area | Capabilities |
|---|---|
| Strategy & Governance | AI Roadmapping · Build-vs-Buy Analysis · AI Cost Management (FinOps) · Responsible AI in Regulated Healthcare (HIPAA) |
| Team Leadership | Building and Growing ML Teams · Technical Hiring and Career Development · Applied Research Management · Executive Communication |
| Technical Depth | Self-Supervised Learning (JEPA) · Video and Multimodal AI · Efficient and Edge AI · Data and MLOps Pipelines (PyTorch, Hydra, Lance, AWS) |