Kindly note that eligibility is limited to candidates residing in Romania.
About us
πππ² πππ₯ππ§ππ¬, founded in 2020, is a recruitment company specialized in IT Recruitment, ππ ππ«ππ’π§π’π§π , but also covering Finance & Accounting, Marketing, HR, and Customer Support roles. We connect clients with top talent across all seniority levels, from juniors to managers, in technologies like Java, Python, .NET, SAP, AI and more. With a dedicated team, flat & success-fee model, transparent pricing, and long-term affordable partnerships ( with relevant volume discounts) , we are committed to delivering guaranteed hires and sustainable recruitment solutions.
Offer
Remuneration & work-life balance:
Competitive salary and trust-based working hours.
Private health insurance.
Generous training budget.
2 extraordinary team events (4 days) per year.
Meal benefit.
Communication & trust:
Open, honest and direct communication. Your ideas are welcome!
A feedback meeting every quarter to help us grow together.
We encourage innovation and are open to new ideas that push the boundaries.
Modern working:
Everything you need for your daily work: MacBook, monitor, headphones and more.
Individual training:
One training day per month and a generous training budget for your personal development.
Buddy program:
An experienced team member will support you from day one to help you get started.
Role
Build decisioning and optimization models that measurably improve how we allocate budgets, choose placements, prioritize experiments, and personalize user communication β while meeting production-grade engineering standards (reliable pipelines, monitoring, and safe deployment). This role exists to turn data into automated or semi-automated actions across marketing and adjacent business functions.
Your Tasks & Responsibilities
- Build prediction models (e.g., conversion/CPA forecasts, propensity/LTV, churn risk, creative or placement performance signals) with strong validation and calibration.
- Build decision policies: rules + constrained optimization (and bandits where appropriate) for allocation, placements, pacing, creative rotation, and personalization.
- Design and analyze experiments/incrementality tests to validate models and policies (not just offline metrics).
- Productionize with βMLE-qualityβ: reproducible pipelines, versioning, monitoring, alerting, and safe rollback; partner closely with Engineering/Data.
- Translate outputs into decision-ready guidance and clear trade-offs (expected impact, uncertainty, constraints).
- Maintain concise documentation of models/policies, data dependencies, and decision logic.
Your goals in this role
- Within 1 month: Deliver a first working βDecisioning MVPβ (one high-impact optimization use case) that produces actionable recommendations on a weekly cadence, with documented assumptions and evaluation.
- Within 2 months: Have 2 decision models/rule systems running in production-like operation (scheduled, monitored, versioned), influencing real decisions (budget/placement/creative/personalization).
- Within 3β4 months: Launch a personalization or allocation model that is validated via experimentation (holdout/A-B) and shows measurable lift vs. baseline policy.
- Within 6 months: Operate a repeatable pipeline for continuous learning (re-training/refresh + monitoring + guardrails) and ship at least 4 high-impact decisioning/optimization improvements used by teams.
Your Key Competencies
Must have
- Strong applied ML + statistics (modeling, evaluation, calibration, leakage prevention, uncertainty-aware thinking).
- Experience building decisioning/optimization (constraints, policies, experimentation-driven iteration; bandits a plus).
- Solid engineering fundamentals: clean code, testing mindset, reproducible pipelines, monitoring/alerting, and secure handling of data/secrets.
- Strong data skills: SQL + Python, feature engineering, data QA, and working with warehouses/lakes.
- Ability to run end-to-end delivery: problem framing β model/policy β validation β rollout β measurement.
Nice to have:
- Ads/marketing domain (DSPs/walled gardens, attribution/MMP concepts) and/or personalization (CRM/push).
- Causal inference / uplift modeling experience.
Your impact in this role
- Higher ROI and fewer wrong moves through evidence-based allocation and optimization.
- Faster learning cycles via experimentation-backed model iteration.
- Reduced manual decision work through reliable decisioning systems and guardrails.
- A scalable foundation for personalization and automation beyond marketing.
Your soft skills
- High ownership: ships, measures, iteratesβdoesnβt stop at βanalysis doneβ.
- Strong stakeholder communication: explains trade-offs, uncertainty, and rollout risks clearly.
- Pragmatic prioritization under ambiguity; focuses on highest ROI levers first.
- Calm, structured execution (production-quality over βclever hacksβ).
- Collaborative mindset with Engineering/Data/Marketing/Finance.
Your Background β Education & Experience
- Masters Degree in Computer Science, Data Science, Statistics, Mathematics, Econometrics, or Business Informatics (or equivalent proven skill).
- 5β10+ years experience in Data Science / Applied ML, with at least one example of models/policies used in production.
- Proven experience working with engineers on production constraints (monitoring, reliability, safe deployment).
Collaboration: CIM only
Location: Romania, full remote
Employment eligibility to work in Romania is required as the company will not pursue visa sponsorship for these positions.
Explore our latest job opportunities on www.key-talents.com or follow our LinkedIn page https://www.linkedin.com/company/key-talents/
#newteam #datascience #keytalents