Mid-Senior Frontier Models and Real-Time Decision Systems Researcher

About the Role

We are seeking a Mid-Senior Frontier Models and Real-Time Decision Systems Researcher to join our team at Sentiro Partners. This role focuses on post-training and offers a unique opportunity to work with frontier-scale models deployed into live, high-stakes systems. As a researcher, you will contribute to machine learning research that ships to production from day one. This is not just academic ML; it’s about real-world applications where latency matters and adversaries exist. Your work will have a direct impact on the organization, and you will be at the intersection of foundational research, large-scale training infrastructure, and real-time optimization.

What You'll Do

• Design and train large-scale neural models for sequential decision making, forecasting, and representation learning under noisy, adversarial, and non-stationary conditions.
• Advance training methodologies: large-batch optimization, curriculum strategies, online/continual learning, and reward-driven fine-tuning.
• Build end-to-end pipelines from data ingestion through deployment and live feedback loops.
• Collaborate with systems engineers to maximize throughput, minimize latency, and maintain stability at scale.
• Ship production-grade models that operate in real-time with strict performance guarantees.

Research Areas

You may specialize deeply in one area or span several:

• Large-scale pretraining and representation learning.
• Reinforcement learning and reward modeling.
• Sequence modeling and temporal architectures.
• Optimization, generalization, and training stability at scale.
• Efficient inference, model compression, and deployment.
• Data-centric ML for high-signal, low-latency systems.

Requirements

• Deep hands-on experience training large neural models (PyTorch or equivalent).
• 4-8 years of professional experience.
• PhD in Machine Learning, Data Science, or Computer Science.
• Strong intuition for optimization dynamics, failure modes, and debugging at scale.
• Comfort working close to the metal: GPU utilization, memory tradeoffs, distributed training.
• A bias toward building and owning systems, not just proposing ideas.
• Background from leading ML environments: evidence of internships or similar in FAANG research teams, frontier AI labs, or equivalent.

Nice to Have

• Publication history is respected but not required; true evidence of researching complex problems is valued.
• Experience in real-world feedback loops and end-to-end ownership of ML systems.

What We Offer

• Relocation assistance and visa sponsorship.
• Real-world feedback loops where your models are evaluated continuously against live outcomes.
• End-to-end ownership of research, training, deployment, and iteration.
• Compute and autonomy to pursue ambitious ideas without academic overhead.
• Work with elite peers in a small, senior team operating at the frontier of applied ML.

Location

This role is based on the East Coast of the USA and is being advertised across the US. If you are ready to build ML systems that operate at the frontier, let's talk!