Skip to content

Recapping the AI, Machine Learning and Data Science Meetup — May 2, 2024

We just wrapped up the May ‘24 AI, Machine Learning and Data Science Meetup, and if you missed it or want to revisit it, here’s a recap! In this blog post you’ll find the playback recordings, highlights from the presentations and Q&A, as well as the upcoming Meetup schedule so that you can join us at a future event.

First, Thanks for Voting for Your Favorite Charity!

In lieu of swag, we gave Meetup attendees the opportunity to help guide a $200 donation to charitable causes. The charity that received the highest number of votes this month was AI4ALL which opens doors to artificial intelligence for historically excluded talent through education and mentorship.. We are sending this event’s charitable donation of $200 to AI4ALL on behalf of the Meetup members!

Missed the Meetup? No problem. Here are playbacks and talk abstracts from the event.

Who needs RLHF When You Have SFT?

This talk centers around Reinforcement Learning from Human Feedback, and more importantly, “Why” is it even needed over Supervised Fine-Tuning? We will also understand in easy terms some current open problems in RLHF as far as research in academia is concerned.

Speaker: Srishti Gureja is an ML engineer and researcher broadly interested in two things: ML efficiency techniques, including but not limited to designing algorithms that make maximum use of the hardware at hand, and the alignment in LLMs using literature from RL. She is currently researching better, simpler methods for aligning language models with Eleuther AI and Alex Havrilla from Georgia Tech. her full-time job is as an ML Engineer at Writesonic, a YC-backed startup.


  • Regarding Alignment:, I have a database of customer queries and our support answers. in an email thread format. We tried to fine-tune mistral-7B, but it couldn’t learn the information correctly. How can SFT or RLHF help us?
  • Does SFT generalize to multi-level conversation?
  • What’s the architecture and complexity of the reward model, or is it the same as our main model?
  • Can you suggest any good video links to try SFT and RLHF on sample data sets for the selected models?

Resource links

Develop a Legal Search Application from Scratch using The Milvus Project and DSPy!

Legal practitioners often need to find specific cases and clauses across thousands of dense documents. While traditional keyword-based search techniques are useful, they fail to fully capture semantic content of queries and case files. Vector search engines and large language models provide an intriguing alternative. In this talk, Mert will show you how to build a legal search application using the DSPy framework and the Milvus vector search engine.

Speaker: Mert Bozkir is a self-taught ML Engineer with a passion for Developer Relations.


  • Is it possible to create a legal case summarization if it involves multimodal data?
  • What is your recommendations for an open source vector database that can handle both structured and unstructured data?

Resource links

Making LLMs Safe & Reliable

Large language models show impressive capabilities, but ensuring their safe and reliable deployment remains challenging. This talk will cover evaluation techniques to assess and improve LLM reliability across key vectors like groundedness and faithfulness. It will also explore detecting vulnerabilities to attacks like prompt injection and PII leaks. Attendees will learn how to build custom evaluations tailored to their use cases.

Speaker: Shiv Sakuja is a former Google engineer, and co-founder of Athina AI, an LLM observability and evaluation platform that helps developers safeguard LLMs in production.


  • Can you give an example of a preset eval?
  • How are you setting up a guard rail on user input,? Do you have a text classifier? Multi-level conversation?
  • How do we set custom evals and custom functions?

Resource links

Join the AI, Machine Learning and Data Science Meetup!

The combined membership of the Computer Vision and AI, Machine Learning and Data Science Meetups has grown to over 20,000 members! The goal of the Meetups is to bring together communities of data scientists, machine learning engineers, and open source enthusiasts who want to share and expand their knowledge of AI and complementary technologies. 

Join one of the 12 Meetup locations closest to your timezone.

What’s Next?

Up next on May 8 at 10 AM PT we have four great speakers lined up!

  • To Infer or To Defer: Hazy Oracles in Human+AI Collaboration – Prof Jason Corso – University of Michigan and Chief Scientist at Voxel51
  • From Research to Industry: Bridging Real-World Applications with Anomalib at the CVPR VAND Challenge – Samet Akcay, PhD – AI Research Engineer/Scientist at Intel
  • Learning Robot Perception and Control using Vision with Action – Brent Griffin at Agility Robotics
  • Anomaly Detection with Anomalib and FiftyOne – Jacob Marks, PhD – ML Engineer/Researcher at Voxel51

Register for the Zoom here. You can find a complete schedule of upcoming Meetups on the Voxel51 Events page.

Get Involved!

There are a lot of ways to get involved in the Computer Vision Meetups. Reach out if you identify with any of these:

  • You’d like to speak at an upcoming Meetup
  • You have a physical meeting space in one of the Meetup locations and would like to make it available for a Meetup
  • You’d like to co-organize a Meetup
  • You’d like to co-sponsor a Meetup

Reach out to Meetup co-organizer Jimmy Guerrero on or ping me over LinkedIn to discuss how to get you plugged in.

These Meetups are sponsored by Voxel51, the company behind the open source FiftyOne computer vision toolset. FiftyOne enables data science teams to improve the performance of their computer vision models by helping them curate high quality datasets, evaluate models, find mistakes, visualize embeddings, and get to production faster. It’s easy to get started, in just a few minutes.