Open weights, real stakes
Battle stories from engineers who have shipped with Open Weights/Sovereign AI models
Accepting submissions
Not accepting submissions
Kanav Dwevedi How far can you get without fine-tuning? Model Quality as architectureDescribe your session (2 paragraphs) This session is a deep dive on building architecture and guardrails around self-hosted open-weight models to deliver quality. We run a production LLM helpline for Gujarati dairy farmers (phone-line and chat) answering questions on animal health, breeding, milk, and schemes, on open-weight models we host ourselves. more
|
Aniket Paluskar Addressing Elephant in the Room: Feature EngineeringThe Elephant in the Room: Feature Engineering Description more
|
Abhijit Roy Enforcing Scope and Data Sovereignty for Open-Weight Indian ModelsIn this session, we’ll enhance the safety of Sarvam, a sovereign Indian LLM, running on RHOAI / OpenDataHub with NeMo Guardrails through the TrustyAI operator. The idea is simple: Sarvam already knows. When you give it sensitive data, it reasons that the information is private. But knowing is not the same as enforcing. The model behaves differently on different runs, keeps no audit trail, and sti… more
|
Sridhar Pillai From LLM to SLM: Enterprise Fine-Tuning That Keeps Open Models Improving in ProdOpen Models That Keep Getting Better: A Self-Improving Distillation Pipeline on Kubeflow ecosystem more
|
Ritesh Shah Sovereign AI at Scale: Engineering a Secure, Private Inference PlatformEnterprises are currently trapped between the urgent need for AI acceleration and the systemic risks of relying on foreign, public SaaS models. This over-reliance is not just a regulatory hurdle—it is an economic and operational vulnerability that mirrors global supply chain dependencies. As “Shadow AI” usage spreads, teams are often forced to build fragmented, insecure tools that lack governance… more
|
Nilesh Kumar Open Weights in Production: A Practitioner's Inference PlaybookRunning an open-weight model on a laptop is easy; getting it to serve reliably in production - on-prem/on-cloud GPUs - is where most teams stall. This session is a hands-on playbook for that journey: model selection, quantization, choosing a serving stack (vLLM, TensorRT-LLM), batching and KV-cache behaviour, the latency-vs-throughput tradeoff, evals, and routing. I’ll use an open-weight model (N… more
|
Prathamesh Kalamkar Presenter Fine-Tuning an LLM: The Lessons We Learned the Hard WayFine-tuning an LLM looks deceptively simple, but the real challenges often lie beyond the training code. In this talk, I share our experience building Aalap, an open-source LLM for Indian legal tasks. We will look at the choices we made, the experiments that surprised us, and the things that did not work as expected. A practitioner’s account of lessons learned the hard way while fine-tuning a dom… more
|
Hosted by
Jumpstart better data engineering and AI futures
Supported by
Venue host
Thoughtworks is a global technology consultancy that integrates strategy, design and engineering to drive digital innovation.