every investment firm calls itself ai-first. far fewer let ai near the numbers they invest on.
they don't trust it, so it all gets checked by hand.
anyone can rent the model. the hard part is everything underneath: every document and dataset the firm has, organized around how it actually works, with code pulling every number and tracing it to its source.
100 seconds on how we built ours.
Life update: I've joined @kepler_ai_hq!
Years of building production-grade AI agents taught me that "probably right" is good enough for a lot of use cases. But when the stakes are high, you need answers you can trace, verify, and defend.
We've spent a lot of time @RunLLM talking to customers and getting feedback — there are some interesting trends in how people are thinking about AI products.
@profjoeyg and I wrote those lessons down: https://t.co/VWFhimIlf2
We're starting to share more about what we're building @RunLLM — the custom assistant for developer-first tools.
@profjoeyg and I started with an overview of what we're building & why: https://t.co/xwFUQRF6FT
We all know that RAG is the killer application for LLMs but did you know that it doesn't work (out of the box)?
Here are some basic steps needed to make RAG actually work:
https://t.co/eAjwwRaBe3
ML/DataEng friends ... at @aqueducthq we're thinking a lot about visibility in production ML systems, have many questions for those of you managing production machine learning in the field.
Please DM if you'd be up for a quick research call. (Quick before the birdsite ends!)
We just released Aqueduct v0.1! 🎉
We're on a mission to remove the complexity from getting data science & ML in production, and this release is a big step in that direction.
We're also on ProductHunt today: https://t.co/xt61Xmze9e
MLOps tools has been mostly for devops and eng given the previous generation. Data scientist are just like the front end becoming more full stack and now needs rethink how to make data scientist proficient
Why are data scientists spending so much time solving the same engineering problems? I am increasingly convinced that MLOps tools are designed for large engineering teams at tech giants and not the every-day data scientists that need them.
1/ MLOps has become increasingly popular of late as a solution to deploying and managing ML models in the cloud. But we believe MLOps is taking the data science and machine learning community in the wrong direction: https://t.co/K0mmmijIq5
Part 2 of @profjoeyg's reflections on his research career! Last week, he discussed the genesis of the ML revolution in powerful abstractions for building models. This post focuses on what comes after building models — using them in real applications: https://t.co/r2Xnoxz6ql
1/ As some of you know, I recently got tenure at @UCBerkeley. I've been thinking about my research career and the simultaneous evolution of the ML space, and I wanted to share what I've been thinking about: https://t.co/pri969xNuX
Our co-founder @profjoeyg recently got tenure @Cal, and he took some time to write about how his career in ML research has been shaped by the broader trends in the space and the technologies that have enabled the machine learning revolution: https://t.co/iI6CBKHi6s