Zing Coach is the world’s first AI Fitness Personal Trainer, revolutionising the way people approach health and wellness. Four years ago, we set out to replace traditional PT sessions with smart, personalised, motivating guidance powered by AI. We’re not just another app – we’re a movement, on a mission to help everyone reach their full physical and mental potential.

Zing Coach is looking for an AI Engineer who will design and build intelligent AI coaching systems that plan, adapt, and personalize workouts in real time. You will develop AI agents and recommendation engines that combine structured user data, health signals, and fitness science principles to deliver safe, effective, and highly personalized training experiences. This includes building multi-step reasoning pipelines, retrieval systems, evaluation frameworks, monitoring, and guardrails, ensuring our AI coach performs reliably in production and continuously improves over time.

Responsibilities

• Building AI agents that plan workouts, adapt in real time, and coach users - including tool use, multi-step reasoning, memory management, and context pipelines

• Building and improving recommendation systems - exercise selection, weight recommendations, intensity scaling - using structured user data

• Integrating health and recovery signals (Apple Health, body scans, fitness tests) into the coach's decision-making process

• Working with the Fitness Science team to define what "good" looks like for subjective tasks - such as whether a workout fits a user's age, injury history, and goals - then building the evaluation frameworks that enforce those standards end-to-end

• Analyzing where agents fail, identifying patterns in those failures, and addressing the ones that matter most

• Designing monitoring and guardrails so agents don't fail silently in production

We expect that you have

• Strong Python skills

• Solid software engineering fundamentals - you ship production code, not just prototypes

• Experience building AI agents - tool orchestration, multi-step workflows, context management, retrieval systems

• Experience building recommendation or scoring systems on structured data - collaborative filtering, calibration, rule-based systems, or hybrid approaches

• Strong analytical instincts - you can define what "correct" means for a subjective recommendation, design metrics that reflect user experience, and spot when a metric is lying to you

• You've fine-tuned or distilled language models to hit cost and latency targets

• You know how to make retrieval work - embeddings, vector search, and semantic similarity

• Experience with classification and regression on structured data - user health signals are messy and continuous

• Ready to learn a new domain deeply - you'll pick up exercise physiology, recovery science, and training principles on the job

Nice to have skills

• Experience building multi-agent systems

• Background in health, fitness, or wellness

• Experience with active learning or human-in-the-loop workflows

• Experience with time-series analysis or forecasting (fitness data is mostly sequences over time)

• Background in ML ops - model versioning, serving infrastructure, pipeline orchestration

Why working at Zing is awesome

• Be part of the fastest-growing fitness and lifestyle startup

• Help shape the internal engine of one of the most exciting AI consumer product

• Excellent work environment: the company is small enough to be person-oriented

• Remote-friendly environment with the option to collaborate in-person

Come be a part of something extraordinary at Zing AI Coach – where innovation meets inspiration, and every day brings new opportunities for growth and impact. Join us and help us create an exceptional user experience for Zing Coach!