Log inSign up
Baladhurgesh
20 posts
@baladhurgesh97

Baladhurgesh

@baladhurgesh97
Building AI Agents
baladhurgesh.github.io
Joined March 2024
141
Following
69
Followers
RepliesRepliesRepostsRepostsMediaMedia

Log in or sign up for X

See what’s happening and join the conversation

Continue with phone
or
Log in with username or email
Terms·Privacy·Cookies·Accessibility·Ads Info·© 2026 X Corp.
  • @baladhurgesh97
    Baladhurgesh
    @baladhurgesh97
    Feb 9
    Now, I can’t forget this analogy while explaining PPO, DPO and GRPO to someone :D
    @ProfTomYeh
    Tom Yeh
    @ProfTomYeh
    Jan 21
    I wrote a short story to explain to my students the evolution of PPO, DPO, GRPO, to GDPO (NVIDIA's new paper). 👇 This story is based on my own personal RL journey to become the family chef. 🍳 (when my wife was my girlfriend) 𝗣𝗣𝗢 I wanted to cook a new dish for our next
  • @baladhurgesh97
    Baladhurgesh
    @baladhurgesh97
    Nov 21, 2025
    Most AI: you’re the product. Silo: you own the data, it’s encrypted, and you still get Deep Research + elite models.
    @freysa_ai
    Freysa
    @freysa_ai
    Nov 21, 2025
    SILO UPDATES 1/ Silo upgrades: new web app, encrypted sync, deep research. Silo started on iOS as a private-by-default AI interface. That same interface now runs on any device, with history that stays end-to-end encrypted across phone and browser.
    1
  • @baladhurgesh97
    Baladhurgesh
    @baladhurgesh97
    Oct 28, 2025
    Not every LLM plays nice. StrategyBench is where we find out who's playing the long game. strategy.freysa.ai/leaderboard #FreysaAI
    @freysa_ai
    Freysa
    @freysa_ai
    Oct 28, 2025
    How do LLMs trust, betray, and cooperate under well considered game scenarios? I built StrategyBench to find out. Open benchmark + scenarios, tuned for research and safety work: strategy.freysa.ai
  • @baladhurgesh97
    Baladhurgesh
    @baladhurgesh97
    May 8, 2025
    Had a great time experimenting with pruning experts in the Mixture of Experts (MoE) layers of LLaMA 4 Scout. Explored multilingual inference, visualized routing patterns, and streamlined the model by removing underutilized experts while preserving output quality.
    @cerebral_valley
    Cerebral Valley
    @cerebral_valley
    May 8, 2025
    Replying to @cerebral_valley
    Finalist: Prunewise This project deploys Scout efficiently: Task and Language based Expert analysis and pruning for Llama4 - scout. @baladhurgesh97 & @BarathAnandan7
    Image
    00:00
  • @baladhurgesh97
    Baladhurgesh
    @baladhurgesh97
    May 1, 2025
    @StartupGrind Conference 2025! It was an incredible experience!
    Image
    Image
    Image
    Image
Advertisement
Advertisement