Saksham

Sounds fun!

Sorry, this isn’t something I was implying. I feel academia can also utilise some of the industry practices. I am sure that might be extremely challenging or some labs might already be working in that way.

Interesting. In the industry for me, it’s not only close mentorship but the requirement to write good code for production. Also the reviews from peers and the need to do certain things through well maintained pipelines could lead to good programming fundamentals.

After my fellowship I have also honed my engineering skills to deploy and train custom LLMs for production use-cases.

Personalized action suggestions in low-code automation platforms Automation platforms aim to automate repetitive tasks using workflows, which start with a trigger and then perform a series of actions. However, with many possible actions, the user has to search for ...

I have a publication at @icseconf.bsky.social I worked on personalizing decoder models with a single token. I also worked on a benchmark to evaluate how well LLMs comprehend code
arxiv.org/abs/2305.10530
huggingface.co/datasets/Hig...

I am applying to Ph.D. programs this cycle with the goal of personalizing LLMs and helping developers write amazing (correct) code. In case you are hiring, please look at my apps!
I have worked as a pre-doctoral fellow at the program synthesis lab @microsoft.com

I think I’ll also make a hitchikers guide to running LMs on ephemeral compute and running it well. Maybe once I have more time ✅

Next step is to serve this under 1 second

Mitski on repeat

Okay ngl. The starter pack list is kinda getting annoying

I have always been confused about what grounding actually means.

But why do I keep opening twitter again and again. Muscle memory needs to be refined

Okay