About this event
Join this online San Francisco AI Meetup on Feb 5, 2026, 9-11 AM PT for four talks on AI, ML and computer vision, with practical demos and live Q&A. This virtual gathering is designed for data scientists, ML engineers, and researchers who want to explore cutting-edge topics in AI and visual intelligence while connecting with a global community.
What to expect:
- Four expert-led talks spanning visual anomaly detection, data-centric pretraining for speech-language models, synthetic data pipelines, and TensorRT optimization for CV workloads.
- Real-world applications and actionable takeaways you can apply to manufacturing, healthcare imaging, quality control, or any ML/CV project.
- A live Q&A session with speakers, plus opportunities to network with fellow practitioners after the talks.
Talk lineup (as described on the event page):
- Unlocking Visual Anomaly Detection: Navigating Challenges and Pioneering with Vision-Language Models. Focused on VAD fundamentals, challenges like data scarcity and domain shifts, and how vision-language models can enable explainable, threshold-free detection. This talk targets researchers and practitioners working on AI-driven inspection and multimodal reasoning.
- Data-Centric Lessons To Improve Speech-Language Pretraining. Examines how data processing, curation, and synthetic data influence SpeechLM performance. The presenters outline practical data strategies to improve pretraining results for spoken language models.
- A Practical Pipeline for Synthetic Data with Nano Banana Pro + FiftyOne. A hands-on walkthrough for generating targeted synthetic data, enriching it with metadata, and managing it in FiftyOne to close real vision gaps and boost model robustness.
- Making Computer Vision Models Faster: An Introduction to TensorRT Optimization. An practical guide to accelerating CV inference, covering layer fusion, precision calibration, memory management, and when to apply TensorRT for production-ready performance.
About the speakers:
- Hossein Kashiani, PhD student at Clemson, focuses on generalizable and trustworthy AI with applications across anomaly detection and multimodal vision.
- Vishaal Udandarao, PhD student (ELLIS) researching foundation models and their generalisation in vision-language contexts.
- Adonai Vera, ML Engineer & DevRel at Voxel51, with extensive CV experience and a passion for open, production-ready AI.
- Tushar Gadhiya, Technical Lead in deep learning and CV, with a track record in academia and indu

