AI Researcher – Multimodal Physiological Modeling

Location: San Francisco Bay Area

Company: KOS AI

Role Type: Full Time

Start Date: January 2025

About Us

KOS AI is developing the next generation of non-invasive metabolic monitoring technology. Our platform combines multimodal sensing, physiological modeling, and advanced machine learning to estimate internal metabolic states continuously and in real time. The work sits at the intersection of biomedical signal processing, artificial intelligence, human physiology, and embedded inference.

This role is ideal for a researcher who enjoys solving scientific problems that have no existing solutions and who is motivated by the opportunity to create entirely new algorithms that push the limits of what wearable technology can measure.

Role Overview

We are seeking an Artificial Intelligence Researcher with strong foundations in time-series modeling, signal processing, and learning from complex physiological data. You will help design algorithms that understand patterns in multi-channel biosignals, model dynamic biological processes, and translate raw sensor activity into meaningful insights.

Your work will directly influence the core intelligence layer of our system and will involve research, experimentation, and the design of new approaches that combine domain knowledge with state-of-the-art machine learning.

What You Will Work On

Your responsibilities will include a selection of the following focus areas, depending on your background:

1. Multimodal Signal Processing and Feature Discovery

• Development of new feature extraction methods for optical signals, motion signals, and physiological biomarkers
• Design of noise-resilient, motion-resilient, and artifact-robust representations
• Exploration of frequency-based, time-domain, and hybrid feature spaces
• Discovery of patterns that reflect underlying biological processes

2. Machine Learning for Physiological Systems

• Research and implementation of models capable of learning from complex, noisy, and high-variability biological data
• Development of ensemble systems, temporal models, and context-aware learning frameworks
• Design of robust inference pipelines capable of running efficiently on constrained environments
• Integration of uncertainty, confidence scoring, and adaptive filtering

3. Biological and Metabolic Dynamics Modeling

• Modeling of dynamic processes related to human physiology, metabolism, and behavioral context
• Exploration of relationships between physiological signals, daily patterns, and state transitions
• Development of personalized data adaptation methods that learn individual physiological baselines and trends

4. Pattern Recognition and Time-Series Forecasting

• Research on forecasting trends and physiological trajectories
• Identification of behavioral and biological events from multimodal data
• Development of systems that adapt to real-world, irregular, and incomplete signals

5. Experimental Research and Dataset Development

• Working with diverse datasets collected across multiple conditions
• Data cleaning, harmonization, and statistical evaluation
• Designing experiments to validate new features, models, and interpretation techniques

What We Are Looking For

Required Qualifications

• Strong foundations in machine learning, time-series modeling, or signal processing
• High proficiency in Python and modern deep learning frameworks
• Experience working with noisy, real-world, or multimodal data
• Ability to explore scientific literature and convert insights into working algorithms
• Curiosity, creativity, and a strong desire to work on difficult open-ended problems

Preferred Qualifications

• Research experience with biomedical signals or physiological modeling
• Background in optical sensing, human physiology, or biosignal analytics
• Experience working with low-power or resource-constrained inference environments
• Familiarity with probabilistic modeling, uncertainty quantification, or ensemble systems
• Demonstrated ability to design and carry out empirical research

Why This Role Is Unique

• You will develop algorithms for a real medical-grade system used by real people.
• You will work on scientific challenges that have no established roadmap and no existing answers.
• You will shape an entirely new category of sensing technology and long-term metabolic analytics.
• You will join a small, highly driven team where every line of research has direct product impact.

Equal Opportunity

KOS is committed to creating a diverse and inclusive workplace. We are an equal opportunity employer and welcome applications from all qualified candidates regardless of race, color, religion, sex, national origin, age, disability, or any other protected characteristic.