Lakshya A Agrawal
@LakshyAAAgrawal
Followers
636
Following
2K
Media
32
Statuses
666
AI PhD @ UC Berkeley | Past: AI4Code Research Fellow @MSFTResearch | Summer @EPFL | Maintainer of https://t.co/LQamOaRksn | Hobbyist Saxophonist
India
Joined December 2013
Real world AI pipelines are often compound, multi-module, and multi-step programs—unlike most RL/GRPO implementations today which optimize a single agent. 🚨 Super excited to release dspy.GRPO, which lets you GRPO tune any arbitrary multi-module, multi-step DSPy program, with.
So many things in the run-up to DSPy 3. Here's a first, EXPERIMENTAL one:. 🚨We're releasing dspy.GRPO, an online RL optimizer for DSPy programs. Your DSPy code as-is can be dspy.GRPO'ed. Yes, even compound multi-module programs. Led by @NoahZiems @LakshyAAAgrawal @dilarafsoylu.
0
15
60
RT @allen_ai: Introducing FlexOlmo, a new paradigm for language model training that enables the co-development of AI through data collabora….
0
47
0
RT @notkevinfarhat: The bottleneck in AI isn't just compute - it's access to diverse, high-quality data, much of which is locked away due t….
0
26
0
RT @luisgschroeder: What if we can guarantee the correctness of LLM responses? Turns out—we can!. We built vCache, the first semantic cache….
0
4
0
I often face this issue where I want to use `pip freeze` to replicate my environment, but instead of wanting to list every single package, only list the minimal set of packages (the ones I likely installed), and fix their versions. `pip list --not-required` solves exactly that!.
5
2
33
RT @NovaSkyAI: ✨Release: We upgraded SkyRL into a highly-modular, performant RL framework for training LLMs. We prioritized modularity—easi….
0
43
0
RT @matei_zaharia: Excited to launch Agent Bricks, a new way to build auto-optimized agents on your tasks. Agent Bricks uniquely takes a *d….
0
45
0
RT @michaelryan207: New #ACL2025NLP Paper! 🎉. Curious what AI thinks about YOU?. We interact with AI every day, offering all kinds of feedb….
0
36
0
RT @AlexGDimakis: I'm excited to announce what we have been working on for months. Announcing OpenThinker3, the strongest 7B reasoning mod….
0
26
0
RT @ShadajL: Finally got to work on a post-graduation side project! Language learning apps have poor support for Indic languages; introduci….
0
6
0
Overheard at SF Agents Meetup: "I wish DSPy added support for GRPO!". Here you go:
So many things in the run-up to DSPy 3. Here's a first, EXPERIMENTAL one:. 🚨We're releasing dspy.GRPO, an online RL optimizer for DSPy programs. Your DSPy code as-is can be dspy.GRPO'ed. Yes, even compound multi-module programs. Led by @NoahZiems @LakshyAAAgrawal @dilarafsoylu.
0
1
12
RT @krshgl: 5 undergrads, no lab, no supervision, and 1 year of "figuring it out" later, we're hyped to present - "Chronocept". It's the….
0
72
0
RT @AlexGDimakis: I thought of DSPy as a prompt optimization tool. But it can optimize the weights of multi-component AI systems too, inclu….
0
10
0
RT @michaelryan207: Check out CAVA 🍾 our new benchmark for end-to-end voice assistants!. Large Audio Models are the next frontier for AI as….
0
4
0
RT @hopkinxx: I’ve built multiple agents that get 50,000+ DAU. DSPy is insane but people “secretly ” use it like me. I am also setting up c….
0
9
0