Skip to content

Deep Dive into Finetuning of LLMs and Startups-AI dev connect

Photo of Pradeep Palani
Hosted By
Pradeep P. and Nihal K.
Deep Dive into Finetuning of LLMs and Startups-AI dev connect

Details

Update: Livestream link - https://meet.google.com/mjv-byhq-qde (starts at 10:30 AM)

Large Language Models are pre-trained to have significant language capabilities but may not have sufficient expertise in a domain, function, company processes, etc. This could limit their usefulness in any setting that requires specific knowledge and capabilities. Finetuning is one way to overcome this and provide LLMs the exposure they need to build expertise in any field or function.

In this session, we will dive deep into the mechanics of finetuning, data requirements, best practices, costs, and business implications, especially in the context of startups and POCs. There will be matchmaking between startups and AI developers interested in finetuning a model together.

Note: This event will be held in-person at Microsoft Reactor Bengaluru (security check and sign-in starts at 10:00 AM) and will also be live streamed (streaming starts at 10:30 AM). Please plan accordingly.

AGENDA

  • Welcome note and introduction to Deep Tech Stars
  • Overview of Microsoft Founders’ Hub
  • Deep Dive into Finetuning of LLMs
    - Intuition around Transfer Learning and Finetuning of images and text
    - LLM overview and architecture for Finetuning LLMs
    - Introduction to PEFT and QLoRA with code walkthroughs
    - Finetuned adaptations (Smart RAG, LLaMA-2, Mistral and XWin-LM)
    - Cost implications and comparisons with semantic embeddings, prompting
    - Q&A
  • POC pitching and sign-up – Startups to pitch their LLM finetuning POC/idea, AI developers to sign up for those project(s) that interest them

SPEAKER
Sai Sundarakrishna, CEO of DeepmAGIc, works with medical fraternity shaping patents, products and research across multiple medical specialties. Before leading DeepmAGIc, a multimodal AI startup, he was the VP of Dream11 Data science (Mumbai), Exec. Director of Wavicle AI products (Montreal) and Senior Researcher/Leader at General Motors and Caterpillar. He is active in Bangalore AI ecosystem, global forums (engg.editorship), conferences, and lectures regularly at several Bangalore meetups on AI. Sai is a Columbia University, NY and Virginia Tech grad in the fields of business, AI and Data Science.
LinkedIn: https://www.linkedin.com/in/sai-sundarakrishna-7b87625

FEE
This meetup is FREE to attend but prior registration is required, with seats being available on a first-come-first-served basis. Please register by filling in the form below.

REGISTRATION
Please fill in this Event Registration Form: https://forms.gle/8E8AV39SfphqYcbY8
Note that RSVPing on meetup.com or in the Meetup app DOES NOT reserve your seat. You must register by filling in the form above, whether you are joining us in person at Microsoft or viewing the livestream online.

Please reach Pradeep at 9066226690 if you need any clarifications or have any challenges in registration. We thank Microsoft Reactor for co-hosting this session with us and look forward to seeing many of you there!

Announcement: If you are interested in assessing your AI skill level, you can join the waitlist for our upcoming product launch (currently open only to those with 2+ years work experience): www.deeptechstars.com

Photo of AI Meetups by Deep Tech Stars group
AI Meetups by Deep Tech Stars
See more events
Microsoft Reactor Bengaluru
Lavelle Road · Bengaluru, Ka