Microsoft launches Frontier Tuning, letting enterprises reinforcement-fine-tune AI models on their own M365 data and workflows.
“With Frontier Tuning, we're making it possible for you to create your own enterprise AI.”
fine-tuning
33 tracked signals on fine-tuning.
Microsoft launches Foundry Managed Compute to run and customize open-source AI models on managed elastic GPU capacity.
“Managed Compute is built to lift that work off your team so the model, not the infrastructure, is what you spend your time on.”
Enterprise teams default to fine-tuning by accident, not conscious architectural choice
“most corporate teams make this decision by accident”
LangChain launches LangSmith fine-tuning in public beta with SmithTune, a CLI to post-train models from agent traces.
“today we're launching LangSmith fine-tuning in public beta with SmithTune, a CLI to allow you to post-train models from your LangSmith traces in one workflow”
Post-training must evolve to let agents adapt to enterprise harnesses without source code access
“there will be these kind of agentic citizens, which you can just deploy once, and they'll be able to adapt to many different types of out of distribution tasks and learn from their interactions”
Microsoft Foundry distills production agent traces into smaller fine-tuned models that run cheaper and faster.
“the question is not just whether my agent works. It is about whether I can afford to run my agent a 100 million times”
Fireworks AI argues open-weight models now match closed frontier models for most enterprise tasks at far lower cost.
“for most enterprise use cases, the gap between openw weight models and frontier closed models has collapsed”
Microsoft Foundry adds reinforcement learning and a low-level training API to turn production agents into cheaper, faster models.
“Think of it as PyTorch as a service.”
Microsoft launches Frontier Tuning: RL-based AI customization within enterprise compliance boundaries
“a new approach to making AI work the way your business does by applying reinforcement learning inside your compliance boundary with your own data, processes, and conventions”
Fine-tuned models for narrow tasks create tech debt that undermines their ROI promise
Amazon Nova Forge enables custom reward functions for multi-turn reinforcement learning via BYOO capability
“A subtly wrong reward can quietly teach the wrong thing while every training curve looks healthy.”
NVIDIA and Prime Intellect Lab simplify Nemotron 3 Nano customization for developers.
AWS's Self-Distilled Reasoning technique prevents catastrophic forgetting during SFT without a teacher model
NVIDIA NeMo Automodel integrates with Hugging Face Diffusers for scaled fine-tuning of video and image models
Use dataset distillation and fine-tuning instead of longer prompts to enforce consistent structured outputs.
“Prompting tells the model what you want right now. Fine-tuning teaches the model a pattern it can follow repeatedly.”
Hugging Face explores fine-tuning methods that may outperform LoRA, the most popular technique.
Open source models still trail closed models on general agentic tasks but can outperform them once fine-tuned for a specific domain.
“because they're open and because you own the weights, you can actually fine-tune or RL them on your specific domain, which can actually make them perform better than those same closed models”
Post-training open-source reasoning models with reinforcement learning keeps agent costs flat while quality improves in Foundry.
“deploying a model into production is just the beginning”
Fine-tuned small open-source reasoning models on Foundry can match frontier models for your domain at a fraction of the cost.
“This is almost like test-driven development for the age of agents.”
As AI products mature, more companies turn to fine-tuning over frontier APIs for performance and cost gains.
“If you tell your LLM to speak like a caveman, you can reduce your tokens by like a lot.”
Multi-turn RL on SageMaker lets small models match frontier reliability for search agents
“Fine-tuning offers a third path: you teach a small model your tools and environment directly. The result is a small model's speed and cost with the reliability that would otherwise require a frontier model.”
NVIDIA fine-tuned Nemotron for Saudi Arabic dialects to address ASR deployment gaps
“Automatic speech recognition must handle how people actually speak, not only the languages and styles that dominate pretraining data.”
Quality beats quantity in SFT: 1,000 curated examples can match models trained on far more data.
“A wrong demonstration gets imitated.”
NVIDIA NeMo AutoModel accelerates fine-tuning of Hugging Face Transformers models.
Hugging Face launches a recurring live tutorial series on fine-tuning coding agents for continual learning.
“I think you could come from from zero here.”
Hugging Face kicks off the 'Build Small' hackathon celebrating small, fine-tunable models over large API providers.
Fireworks AI's managed inference and training service is now available on Microsoft Azure.
“Fireworks is a managed service where we do both AI inference performance as well as AI training all as a managed service that you can now use on Microsoft Azure.”
Combining SFT and DPO on SageMaker AI improves a small model's tool-calling accuracy for agents.
“When an agent picks the wrong tool, formats parameters incorrectly, or breaks a workflow chain, task completion times grow, error rates rise, support costs increase, and user experiences degrade.”
uniopen fine-tuned Amazon Nova 2 Lite for proprietary two-axis retail content moderation in production
AWS recommends ~2,000 high-quality samples as baseline for supervised fine-tuning tasks
“plan for roughly 2,000 high-quality training samples for a typical SFT task”
Hugging Face is teaching GRPO reinforcement learning for agent training in a live stream series
“it's pretty straightforward to learn from the available options and you can apply it on most use cases”
AWS details hyperparameter tuning practices for customizing frontier models on Amazon Nova Forge.
“If any of them are wrong, you trade one problem for another.”
Hugging Face releases guidance on training multi-vector embedding models with Sentence Transformers