ML Engineer - Tabular Data & Experimentation

Location: Warsaw, Poland. Hybrid schedule with 4 days in the office and 1 day working from home.

Company

hirify.global is a live-streaming B2C platform with more than 450 million registered users, enabling creators to produce live content, engage with fans, and monetize their talents.

What you will do

• Build end-to-end machine-learning solutions for tabular and panel data, including feature engineering, time-aware validation, gradient boosting, calibration, and drift monitoring.
• Design A/B tests covering hypotheses, statistical power, minimum detectable effects, multiple testing, peeking, variance reduction, interference, and network effects.
• Use causal inference methods such as difference-in-differences, synthetic controls, instrumental variables, and uplift modeling when online experimentation is not feasible.
• Design rigorous offline and online evaluations for LLM-based components, including metrics, proxies, drift detection, power analysis, and guardrails.
• Translate product and business questions into ML formulations, metrics, losses, constraints, and practical trade-offs.

Requirements

• 5+ years of applied machine-learning experience with measurable product impact.
• Deep knowledge of tabular and panel data, temporal leakage, non-stationarity, and differences from i.i.d. data.
• Strong statistics, experimentation, and causal-inference skills, including selecting and defending appropriate methods.
• Production-grade Python development and production-analytics SQL, including complex joins, window functions, and query-plan reasoning.
• Ability to evaluate when LLMs, gradient-boosted decision trees, or classical methods are appropriate, including their cost and performance trade-offs.

Nice to have

• Experience building or substantially reworking production recommender systems, including candidate generation, ranking, reranking, or two-tower architectures.
• Experience with LLM-augmented retrieval and ranking, semantic retrieval, embeddings, or LLM rerankers.
• Experience with DoubleML, EconML, MLflow, Weights & Biases, or feature stores.
• Knowledge of LLM-as-judge methodology and empirical prompt optimization.

Culture & Benefits

• Stock options grant from a Silicon Valley company.
• Competitive salary.
• Medical insurance for the employee and 75% coverage for relatives.
• Lunch budget, parking, and a Multisport card.
• On-site office environment with a collaborative and energetic team atmosphere.