Join us for an engaging ML Training and Orchestration Meetup, where we dive into the art and science of building scalable, efficient workflows for machine learning models.
As the complexity of AI systems grows, so does the need for streamlined processes that can handle diverse data pipelines, distributed training, and seamless deployment. This meetup is a unique opportunity to share knowledge, exchange ideas, and learn from experts who are shaping the future of ML orchestration.
We’re inviting speakers with deep insights into cutting-edge tools and strategies, offering attendees a chance to explore best practices and emerging trends. Whether you’re an ML practitioner, a DevOps enthusiast, or just curious about the field, come and be part of the conversation driving innovation in this critical domain!
Arnab is a software engineer on the Mast team at Meta. He is very interested in working on distributed systems with a focus on ML infrastructure. Before Mast, Arnab was a manager in the Ads Infrastructure group at Meta and in a previous life, spent a decade working at Microsoft on Windows Phone and Xbox and a few years at Tableau working on a runtime framework for data visualization. In his spare time, Arnab likes to spend time with his family and also dabbles in photography on the side.
Snehal is a SWE manager at Meta. She supports the data loading and preprocessing teams for AI training. Before AI-Training, Before AI training, Snehal supported the Tier-0 caching services that serve Meta’s apps and online services to billions of users. Prior to Meta, Snehal built the database product which was the highest revenue stream at the data management startup, Rubrik. In her spare time, Snehal loves to be outdoors hiking, biking, and gardening.
Muthu is a software engineer at Meta working on problems related to the scheduling of distributed model training workloads. His key accomplishments prior to this include being a founding member of ZippyDB (one of the world’s largest, distributed, NoSQL store), Akkio (a data locality solution that works at Meta scale) and RIM (a general purpose resource management framework core to many of Meta’s pivotal infrastructural building blocks). Prior to his 12-year career at Meta, Muthu worked at Microsoft solving a variety of distributed systems problems.
Haowen Ning is a seasoned engineering leader with 13+ years of experience in AI infrastructure, large-scale systems, and mobile development. At LinkedIn, he leads the AI training platform, driving rapid model experimentation and LLM training infrastructure. Previously, he led AI modeling and smart messaging features for LinkedIn Messaging and spearheaded key Azure SQL projects at Microsoft. Haowen also created AnyMemo, a popular open-source Android app with 250,000+ downloads. He excels in AI platforms, cloud systems, and mobile solutions.
Biao He is an experienced software engineer with a strong focus on AI platforms and machine learning infrastructure, bringing over 8 years of expertise in ML model training and scalable systems. As a Staff Software Engineer at LinkedIn, Biao drives advancements in ML model training platforms to support high-performance AI workflows.
Previously, Biao contributed to knowledge mining and ML platform development at Microsoft, enhancing machine learning capabilities for enterprise applications. At Zynga, he honed his software engineering skills while working on complex systems in the gaming industry.
Biao’s expertise spans ML infrastructure, model training, and AI platforms, with a proven ability to deliver impactful solutions at scale.
Chaoying is a senior software engineer at Netflix on the Model Development and Management team. Her work focuses on accelerating the productivity of ML practitioners by designing simple and consistent abstractions for core ML pipeline properties. Prior to Netflix, she contributed to recommendation ML infrastructure at Meta, specializing in online training and distributed inference.
Darin Yu is a senior Software Engineer with a robust background in data engineering and machine learning infrastructure. Currently, Darin is part of the Model Development team at Netflix, where he focuses on enhancing the open-source ML orchestration library, Metaflow, originally developed by Netflix. Prior to his role at Netflix, Darin contributed to YouTube Ads, where he concentrated on conversion measurement, employing ML-inspired solutions to enhance both coverage and accuracy. Darin's expertise lies in developing scalable ML systems and crafting innovative solutions that drive impactful results in data-driven environments.
Shashank is a software engineer in the Machine Learning Infrastructure team at Netflix. He is a core contributor of Metaflow, an Open Source framework for orchestrating machine learning and data science pipelines. Prior to Netflix, he pursued research in robotics and computer vision, with publications at several top conferences like IEEE IROS, and AAAI ICWSM.
Yubo Wang is a dynamic software engineer with expertise in AI training platforms, distributed systems, and traffic management platforms. Currently, as a Senior Software Engineer at LinkedIn, Yubo is instrumental in building the company's AI Training Platform, including introducing and productionizing Flyte to support multi-purpose training for Ranking, RecSys, and LLMs.
Previously at Hulu, Yubo contributed to the Traffic Management Platform, developing an API Gateway with Envoy Proxy and internal service mesh solutions using Kubernetes and Istio. His work on Hulu's load testing framework enabled support for 3.3M RPS.
Parking, Check-in, and Important Info
We are excited to have you at Meta. Our office has safety policies and requirements you should be aware of prior to your visit. Here's what you need to know:
Photos, Videos, and Recording Devices:
Visitors may take photos or videos outside, in the cafes, in the lobbies and in front of
the living Meta Wall just inside the main entry points. Visitors are NOT allowed to
take pictures or videos in the following areas:
Inside Meta workspaces
Inside Meta conference rooms
Near Meta laptops
All Meta whiteboards
Driving instructions:
Address: 1 Hacker Way, Menlo Park, CA 94025
From CA-84 West / Decoto Road:
Turn onto Hacker Way
From CA-84 West/ Decoto Road right onto 1 Hacker Way
From CA-84 East/ Menlo Park turn left onto 1 Hacker Way
Stay in the right lane to make a slight right
Continue down road following signs for MPK 15
Park in any open spot at MPK 15
Do not park in any assigned parking (ADA, Vanpool, EV).
Arriving via Rideshare:
Taxi, Uber, Lyft, and other ride hailing services can only pick up and drop off guests at buildings MPK 20/21.
MPK 20/21 is across the highway from the check-in building of MPK 15.
Free shuttle service will be provided from the Uber/Lyft dropoff location to MPK15. ML Meetup signs will be posted on the shuttle.
Check-in Process:
Upon arrival, check-in using your email address and sign our NDA via the tablet in the lobby.
Sign in under “Event” or “Vendor” on the tablet or input your 6 digit code sent via email.
Provide reception staff with a form of government ID. Photos of IDs will not be accepted.
You will receive a name tag on a green lanyard immediately after check-in.
Ensure you are wearing your lanyard around your neck at all times during your visit at Meta.
You will be escorted by a Meta employee from the lobby to the event space