Jacob Sanders

@jacosand

Machine learning, artificial intelligence, neural networks, active learning. Machine Learning Engineer @Osmo_Labs. 🎓 Ph.D. @HarvardGSAS, A.B. @Harvard.

New York, NY Joined July 2014

226 Following 156 Followers 93 Posts

Latest posts 5

Jacob Sanders @jacosand Sep 10

Over the last several weekends, I built my own end-to-end language model from scratch using CS336 from @stanfordnlp.

Though I've seen attention and transformers more times than I can count, building the entire system with no details spared was a very different experience. 🧵

Jacob Sanders @jacosand Jun 9, 2024

I moved to the West Village in NYC in April, and by now I'm more or less settled in! Who's around NYC and wants to get together?

Jacob Sanders @jacosand May 28, 2024

I'll be in Boston this coming weekend (Friday, May 31 to Sunday, June 2) for my Harvard reunion, and would love to see people, even if not part of the reunion! Who's around? I'll be there all day Sunday after the reunion ends, and can also possibly make time on other days.

Jacob Sanders @jacosand Jun 13, 2023

Proud to be part of this massive #llmhack manuscript invited by none other than my Ph.D. advisor @A_Aspuru_Guzik, and to be part of the pioneering movement to incorporate LLMs into how we solve chemical problems!

Kevin M Jablonka @kmjablonka Jun 13, 2023

11. @QaiAlex, @jacosand, @Jiale_Shi, Stefan Bringuier, @BrendenPelkie, Marcus Schwarting use LLMs to convert procedure texts in the ORD to structured JSON

github.com/qai222/LLM_org…
x.com/QaiAlex/status…
View quoted post on X

Jacob Sanders @jacosand Mar 30, 2023

My first foray into fine-tuning LLMs to convert textual descriptions of organic procedures into structured JSON data suitable for downstream ML tasks. Was exciting to work with @QaiAlex and the rest of the team at #llmhack and fine-tune the LLM!

Qianxiang Ai @QaiAlex Mar 30, 2023

Use LLM to parse free text synthesis recipes to structured data! With only 300 training pairs, the fine tuned model can already pick up chemical identities/amounts and generate valid JSON in ORD schema. Check out our video demo!

View quoted post on X