Pablo Castro

Its a super feeling to be able to interface a Streamlit UI to a Google ADK Agent, and deploy on GCP Cloud run. Imp Learnings : how to use the Streamlit session state and interface with Agent Memory (surprisingly, this intricate code is developed by Gemini code-assist!) how to flush out the session state and start a new Chat when a specific keyword is typed how to build a docker image on the Google Cloud (no local build) and deploy it Checking Cloud run logs and taking necessary action Auto-scaling and instance configuration of Cloud run Some quick improvements possible : summarizing the chat context when it gets too big (already done this in a previous post), and of Course UI with nice look and feel. Finally it feels great when we see the UI running on Cloud and interacting with the ADK Agent. Ref: https://lnkd.in/gqZEN8Hw #streamlit #gcp #adk

14

1 Comment

A new paper from SJTU and GAIR (daVinci-Env) makes a compelling case that we've been looking at the wrong variable in SWE agent training. The premise is refreshingly honest: the open community doesn't lack clever RL algorithms or agent architectures. What it lacks is a large-scale supply of executable, verifiable training environments. So they built one. 45,320 Docker environments across 12.8k Python repositories, synthesized from 572K GitHub PRs through a multi-agent builder pipeline. The environments come with Dockerfiles, evaluation scripts, and distributed construction infra - all open-sourced. The engineering decisions matter more than the headline numbers: -> Builder decomposed into retrieval / setup / evaluation / analysis roles (not one monolithic agent) -> Dual-run validation protocol: test-only must fail, test-with-fix must pass -> Structured failure routing that becomes free process supervision for training -> Success-frequency curriculum: keep instances where the model succeeds 1-2 out of 4 attempts The most convincing result is the environment-source ablation: holding model and scaffold constant, swapping only the training data source, OpenSWE beats SWE-rebench in all four settings - up to +12.2 points. That's direct evidence that environment quality is a first-order training variable. Where the paper is weaker: the "difficulty-aware curation" story lacks clean ablation, the 32B headline slightly oversells relative to the strongest RL baseline (+1.0, not +4.6), and whether the curated trajectories are fully released remains unclear. The real takeaway for anyone building post-training pipelines: verifiable feedback loops and environment quality may matter more than the choice of optimization algorithm. This paper has quietly assembled every ingredient needed for an RL-for-code environment engine - real repos, executable sandboxes, verifiable outcomes, structured error attribution. The next step is online policy improvement on top of that substrate. https://lnkd.in/e2cFM-zu

5