Sr. Applied Behavioral Scientist

Job Description

Job Description

The Company:

VeSync is a portfolio company with brands that cover different categories of health & wellness products. We wouldn’t be surprised if you have one of our Levoit air purifiers in your living room or a COSORI air fryer whipping up healthy and delicious meals for you every night.

We’re a young and energetic company, we’ve had tremendous success, and we are constantly growing our team. As we garner more industry attention – just check out our accomplishments and awards by CES Innovation, iF Design, IGA, and Red Dot – we also need driven and talented people to join our team.

That brings us to you, and what you’ll be joining. Our teams are smart and diligent and take ownership of their work – they’re confident in their work but know how to collaborate with open ears and a spirit of learning. If you’re down-to-earth, approachable, and easy to strike up a conversation with, this may be a great fit for you.

Check out our brands:

levoit.com \u007C cosori.com \u007C etekcity.com

The Opportunity:

We are seeking a Senior Applied Behavioral Scientist to join our growing US-based Behavioral Science team as the senior technical lead for key components of our system. This role combines two functions that are closely intertwined in our product: owning the decision logic that determines what the system does for a user, and owning the statistical and causal-inference architecture that determines whether an intervention is working for that user.

On the decision logic side, you will lead development of our just-in-time adaptive intervention design system in partnership with the AI Team — a core system that determines when and how the app engages users, guiding them at moments that matter most. On the causal inference side, you will own the personalized learning loop that determines intervention effectiveness at the individual level — specifying the attribution scheme, defining instrumentation requirements for Engineering, and partnering with ML on the recommendation and causal-inference architecture. You will also have the opportunity to publish.

This is a high-leverage, intellectually demanding role for a senior behavioral scientist who combines data science-level rigor with behavior change theory and hands-on expertise in building decisioning systems, collecting and working with behavioral data, causal inference and adaptive experimentation, and who is comfortable translating both into practical, sprint-compatible infrastructure at an industry pace.

Just-in-time Adaptive Intervention System

• Support Behavioral Taxonomy Development: Design and build a structured behavioral taxonomy in close collaboration with the AI Team. Define the taxonomy of behavioral targets, barrier profiles, BCT mappings, and intervention modalities. Ensure the taxonomy is structured for machine readability and downstream use in the recommendation system.
• Translate Academic Frameworks into Applied Schema: Convert the BCTTv1, COM-B framework, and relevant behavior change evidence into a practical classification system that Engineering and ML teams can operationalize.
• Review and Validate Behavioral Logic and Algorithms: Provide expert review of intervention logic, BCT-to-barrier mappings, and taxonomy edge cases. Flag areas where the system’s behavioral logic deviates from the evidence base.
• Maintain Evidence Standards: Bring a publication-grade standard for evidence evaluation. Help the team distinguish between well-supported, plausible, and speculative behavioral claims within the taxonomy.
• Lead JITAI System Development: Design and build the just-in-time adaptive intervention design system in close collaboration with the AI Team — the decision logic that determines whether, what, and when to prompt a user at each moment of opportunity in the app.
• Define Decision Points & Tailoring Variables: Specify the decision points at which the system evaluates whether to intervene, and the tailoring variables — behavioral state, context, receptivity, and prior response history — that inform each decision.
• Design Decision Rules from Behavioral Evidence: Translate BCTTv1, COM-B, and relevant behavior change evidence into decision rules that map tailoring-variable values to specific intervention options, ensuring every prompt delivered is behaviorally grounded.
• Own the Behavioral Taxonomy & Delivery Constraints: Define and maintain the taxonomy of prompts, nudges, and intervention modalities the decision logic can select from, along with rules for cadence, cooldowns, and sequencing that protect against message fatigue and habituation.
• Review, Validate & Maintain Evidence Standards: Provide expert review of decision rule performance in production, flag areas where real-world behavior deviates from the evidence base, and bring a publication-grade standard for distinguishing well-supported, plausible, and speculative behavioral claims driving the system.

• Own the Learning Loop Attribution Scheme: Design the statistical and causal framework for the Learning Loop before product launch. Specify micro-randomization strategies, off-policy evaluation methods, and individual-level effect estimation approaches that will power personalized intervention delivery.
• Define Instrumentation Requirements: Partner with Engineering to specify the event logging and data infrastructure needed to support the learning loop. Ensure instrumentation is in place before launch to enable attribution and effect estimation.
• Specify Adaptive Trial Designs: Design and advise on adaptive experimentation methodologies including Multi-Arm Bandits (MABs), Micro-Randomized Trials (MRTs), and contextual exploration strategies appropriate for within-person behavioral data.
• Support N-of-1 Experiment Infrastructure: Develop the analytical framework for N-of-1 self-experiments within the product. Define how individual-level effect estimates are computed, communicated, and updated over time.
• Advise on A/B Testing and Factorial Designs: Consult on experiment design for population-level tests. Provide guidance on power, sample size, randomization, and the interplay between individual and group-level inference.

• Partner with the AI/ML Team: Work directly with ML Engineers on the recommendation and intervention system architecture, aligning taxonomy structure and causal-inference methodology with the technical architecture of the system.
• Advise on Methodology: Consult on measurement design, study methodology, and intervention evaluation frameworks as needed. Serve as an internal expert reference for the BeSci team.
• Collaborate with the Behavioral Analyst: Partner with the Behavioral Analyst to ensure the taxonomy and statistical learning framework are aligned with the behavioral theory driving intervention selection.

• Education: PhD in Behavioral Science, Health Psychology, Behavioral Medicine, Data Science, Machine Learning, or a closely related field is required, with a strong publication track record demonstrating expertise in combining behavior change theory with closed-loop system design. Advanced training or applied experience in causal inference, reinforcement learning, or a related quantitative discipline is also required.
• Experience: 5+ years in a senior behavioral science role that combines intervention design or taxonomy/framework development with causal-inference or quantitative research, ideally in health tech, digital health, or algorithmic decision-making (ADM) contexts.
• Behavior Change Frameworks: Deep expertise in BCTTv1, COM-B, and related behavior change models, with prior experience applying these frameworks to digital health or technology-mediated interventions strongly preferred.
• Causal Inference & Adaptive Methods: Demonstrated expertise in causal inference, off-policy evaluation, and adaptive trial designs. Hands-on experience with Micro-Randomized Trials (MRTs), Multi-Arm Bandits (MABs), and Reinforcement Learning (RL) algorithms in applied settings.
• N-of-1 and Sequential Methods: Experience with N-of-1 experimental designs, sequential decision-making, and within-person inference. Comfort with the statistical challenges of small-n, high-frequency behavioral data.
• Applied Comfort: Must be comfortable operating at an industry pace — translating academic rigor into actionable, sprint-compatible deliverables without sacrificing scientific integrity.
• Technical Collaboration: Proven ability to work closely with engineering and ML teams on knowledge representation, taxonomy design, structured data schemas, and data instrumentation requirements.
• Communication: Excellent written and verbal communication. Able to write clear methodology specs, taxonomy specifications, and review memos, and to present complex statistical and behavioral concepts to both technical and non-technical audiences.
• Industry Knowledge: Experience in health tech, behavior change platforms, or consumer wearable/connected device ecosystems is strongly preferred.
• Engagement Type: This is a full-time role. Candidates should be prepared to commit to a defined scope and timeline, with structured touchpoints across the Engineering, ML, and BeSci teams.

• Tustin, CA

• Starting at $200K Annually

• 100% covered Medical/Dental/Vision insurances for employee AND spouse + dependents!
• 401K with 4% employer match (eligible after 90 days of employment) and immediate 100% vesting
• Generous PTO policy + paid holidays
• Life Insurance
• Voluntary Life Insurance
• Disability Insurance
• Critical Illness Coverage
• Accident Insurance
• Healthcare FSA
• Dependent Care FSA
• Travel Assistance Program
• Employee Assistance Program (EAP)
• Fully stocked kitchen