Machine Learning Researcher
Inference.net provides fully managed AI infrastructure for deploying, observing, tracing, training, and evaluating open-source, custom, and fine-tuned AI models at scale. It serves AI-native product teams looking to switch from providers like OpenAI, Anthropic, and Gemini to optimized infrastructure for lower cost and faster performance.
Funding
Projects
About Inference.net
Inference.net offers a suite of infrastructure products for AI-native teams, including Deploy (turn-key, globally distributed model hosting with high uptime), Observe (LLM observability with monitoring, tracing, and debugging tools), Trace (capturing agent LLM calls, tool calls, and framework steps), Train (custom model fine-tuning in days), Evaluate (rigorous benchmarking before production deployment), and HALO (open-source agent optimization). The platform hosts models such as GLM-5.2, Kimi K2.6, MiniMax-M2.5, and GPT-OSS 120B, and offers proprietary open-source models like ClipTagger and Schematron. Clients include engineering teams such as Cal AI and Gravity Ads, who use the platform to cut latency and costs while maintaining frontier-level model quality. The company is SOC 2 Type II compliant, indicating a focus on production-grade, enterprise-ready AI infrastructure.
Skills
About the Role
You will conduct research into experimental models, training systems, and modalities to create novel products for customers. Your work spans exploring new architectures and learning methods to optimizing latency and efficiency, all with the goal of delivering better models to customers. Your north star is pushing the frontier of what's possible in LLM post-training. You'll explore new techniques, run rigorous experiments, and when something works, help bring it into production with your teammates. This includes training models for customers and running evaluations as part of validating your research. You'll report directly to the founding team and will have autonomy, a large compute budget/GPU reservation, and technical support to explore ambitious ideas and ship the ones that work.
Requirements
- 3+ years of experience training AI models using PyTorch
- Deep understanding of transformer architectures, attention mechanisms, and model internals
- Hands-on experience with post-training LLMs using SFT, RLHF, DPO, or other alignment techniques
- Experience with LLM-specific training frameworks (e.g., Hugging Face Transformers, DeepSpeed, Megatron, TRL, or similar)
- Strong experimental methodology, including ability to design, run, and analyze rigorous experiments
- Track record of implementing ideas from recent ML papers
- Experience training on NVIDIA GPUs at scale
- Strong foundation in ML fundamentals: optimization, loss functions, regularization, generalization
Responsibilities
- Research and experiment with new model architectures to improve quality, efficiency, or capability
- Explore methods to decrease inference latency and improve serving efficiency
- Run experiments with new learning methods, including novel approaches to SFT, RLHF, DPO, and other post-training techniques
- Perform reinforcement learning research to improve model alignment and capability
- Develop and improve the distillation pipeline for training high-quality models from frontier teachers
- Train models for clients and run evaluations to validate research findings in production settings
- Create robust benchmarks and evaluation frameworks that ensure custom models match or exceed frontier performance
- Stay current with ML research and identify techniques that can improve the platform
- Collaborate with applied engineers to bring successful research into production systems
- Document findings and share knowledge with the team
Benefits
- Equity
- Comprehensive benefits