Members-Only
Recent Talks & Demos are for members only
You must be an AI Tinkerers active member to view these talks and demos.
json_partial_python: Reliable Structured Outputs
The talk explores using the json_partial_python library to generate reliable structured outputs from small models, improving text-based workflows before Pydantic validation.
leverage json_partial_python library to get reliable unstructured outputs from small models. (3b, 8b)
Rust `json_partial` error-tolerantly parses imperfect LLM JSON, including Markdown extraction.
Rust-backed Python library for resiliently parsing malformed JSON from LLM structured outputs.
Related projects
Making-llm-reliable
Delhi
The session outlines practical methods to improve LLM reliability, covering error reduction, evaluation metrics, and deployment strategies for…
Building AI Workflows
Delhi
This talk covers building complex AI workflows using Julep, an open-source platform that simplifies creating agentic AI applications…
AI in compliance
Pune
Learn how AI agents automatically analyze documents, identify compliance gaps, and provide real‑time monitoring for ISO‑27001/SOC, using LLMs,…
AI in Education
Pune
The talk explains Sahay, an AI platform using RAG and LLMs to deliver personalized learning paths, career counseling,…
LLM Security
Delhi
This talk covers vulnerabilities in agentic systems and LLM-driven applications, including a theoretical overview followed by a demonstration…
AI Decision-Making in Low / No Trainable Data Domains
DC
This talk explores using expert-created rules of thumb to guide large language models in specialized domains with little…