Head of Artificial Intelligence / Machine Learning (Head of AI / ML)
Head of AI / ML is responsible for machine learning and AI in the product: ranking and matching models, anti-fraud, personalization, and the team of ML engineers and researchers.
What a Head of Artificial Intelligence / Machine Learning does
The Head of AI / ML leads everything in the product that runs on machine learning, from recommendations and matching to photo moderation, detection of fake and bot profiles, and churn prediction. In dating, this is one of the main sources of value, because the quality of matching directly determines user retention, conversion to paying users, and the share of suspicious traffic. Companies in high-risk verticals keep this role separate from the general CTO, because it requires both deep engineering expertise and an understanding of regulatory and reputational risks.
This position usually reports to the CTO or CPO and manages ML engineers, data scientists, MLOps specialists, and sometimes a data labeling team or internal tooling team. The Head of AI / ML agrees with product on which tasks are solved by models and which by rules, and ensures that models go into production reliably, explainably, and in compliance with data protection and content moderation requirements.
Responsibilities
- Define the AI strategy and priorities for the quarter and year, linking them to product and revenue goals
- Lead the team of ML engineers, data scientists, and MLOps specialists, distribute tasks, and hire key specialists
- Own the matching and profile ranking models: goal setting, offline and online evaluation, rollout through A/B tests
- Develop anti-fraud models to detect bots, fake profiles, fraudulent messages, and multi-accounts
- Organize automated and semi-automated moderation of photos and texts in line with platform and app store requirements
- Build data and model training pipelines together with data engineers: labeling quality, dataset versioning, drift monitoring
- Agree with lawyers and compliance on the use of personal data, photo storage, and explainability requirements for decisions
- Protect the budget for compute and external APIs, and control inference and cloud resource costs
- Present ML results to leadership and product teams, translating model metrics into business language
Skills
- Deep understanding of machine learning: ranking, classification, recommender systems, time series
- Experience deploying models to production and being accountable for their performance under load
- Ability to frame ML problems in terms of product and revenue, not just model accuracy
- Managing teams of engineers and researchers: hiring, evaluation, development, prioritization
- Understanding of A/B testing and causal analysis, including the limits of online experiments
- Knowledge of anti-fraud risks and adversarial attacks: fraudsters adapt to the model
- Communication skills with lawyers and compliance on personal data and content moderation
- Estimating compute costs and architectural decisions, balancing speed and quality
- Ability to explain complex decisions to non-technical executives without oversimplifying conclusions
Tools
- Python, PyTorch or TensorFlow for model training
- SQL and data warehouses for datasets and analytics
- Cloud ML platforms and MLOps tools for training, deployment, and monitoring of models
- Version control and experiment tracking systems (Git, MLflow or similar)
- A/B testing tools and internal experimentation platforms
- Jira or a similar tracker for team planning
- Data labeling and content moderation systems
How performance is measured
- Growth of key product metrics driven by ML models: match rate, message replies, retention
- Share of traffic filtered by anti-fraud and share of false positives
- Moderation speed and share of content processed without manual review
- Inference and compute cost per active user
- Time from model idea to production
- Production stability of models: number of incidents and deviations from baseline metrics
- Turnover and engagement of the ML team
By vertical
Dating
In dating, the main AI tasks are match quality and fighting fake profiles, bots, and fraud in messaging. The Head of AI / ML works with high sensitivity around personal photo data and with requirements from app platforms and payment partners, so moderation and anti-fraud models take up most of the priorities here.
Levels
Head of / Director
The Head of AI / ML is responsible for AI strategy across the entire product, the budget and team hiring, and the risks that models create for the business and users. Decisions are made at the level of architecture, priorities, and collaboration with the CTO, CPO, and the legal function, rather than individual models.
How to get there
People usually move into this role from an ML engineer or data science lead position, where they already had experience deploying models to production and managing a small team. It is important to have proven experience where a model influenced product metrics, not just offline evaluation. A background in mathematics, statistics, or computer science is essential, but it is not enough without operational experience.
For those who want to reach this role, it helps to progress step by step from ownership of individual models to ownership of a whole area: leading projects with several teams, defending decisions before leadership, and taking part in hiring and evaluation. It is worth learning anti-fraud, content moderation, and regulatory limits on personal data, since in high-risk verticals these determine half of all decisions.
Growth within the area usually comes through expanding scope: from one model to a platform, from one product to several, from technical expertise to managing budget and risk. Public courses that provide the right practical experience are almost nonexistent, so experience is built on the job.
Next steps: CTO or VP Engineering, Chief Data Officer (CDO), VP of AI, Director of Data, Head of Data Science
FAQ
How does a Head of AI / ML differ from an ML engineer?
An ML engineer builds and deploys specific models, while the Head of AI / ML is responsible for the whole area: strategy, priorities, team, and risks. They decide which problems should be solved with models at all and defend those decisions before leadership. Technical depth still matters, but the core work shifts toward management.
Do I need to know programming to hold this position?
Yes, at least at the level of understanding how models, data, and production systems work, even if you no longer write code every day. Without this, it is hard to judge whether a timeline is realistic or to understand why a model is degrading. Most people in this role came from engineering.
Which risks in high-risk verticals such as dating matter for this role?
The main risks relate to personal data, especially photos, to content moderation, and to fraud that adapts to the model. Requirements from platforms and payment partners regarding content and restrictions are also important. The Head of AI / ML must agree on these issues with lawyers and compliance in advance.
What determines income in this position?
Income depends on company size, product scale, the maturity of the ML function, and how strongly models affect revenue and retention. Experience deploying models to production and managing teams also matters, as does the market in which the company operates.
How do I write a job posting for Head of AI / ML?
Describe which product problems the models solve, what stack and data scale are involved, and what authority the role has over budget and hiring. State who the position reports to and which teams it works with. Specify requirements for personal data and moderation, since these significantly narrow the pool of candidates.
Looking for a job in the industry?
Post an anonymous resume in our bot — recruiters from iGaming, payments and crypto search it every day.