01
Data science review & due diligence
Review current models, challenge investment assumptions and identify overlooked opportunities. Leave with clear priorities and a practical roadmap.
Business priorities · technical evidence · next steps
INDEPENDENT ADVISORY · TECHNICAL LEADERSHIP
Your trusted
Data Science
partner.
I specialise in predictive modeling, causal inference and marketing mix modeling. I turn data science and machine learning into better commercial decisions, from model reviews to a practical roadmap.
Data scientist. Technical leader. Hands-on mentor.
01 — INDEPENDENT TECHNICAL ADVISORYSCROLL TO EXPLORE
13+
years in applied
data science
MSc
Mathematics
Mathematics
Specialisation in statistics
and data analysis
Technical
leadership
leadership
Setting research direction
and shaping DS solutions
Technical
mentorship
mentorship
Developing specialists
and strengthening teams
01 WHERE I CAN HELP
Independent expertise.
Practical business decisions.
For businesses improving models, measurement and team capability, and investors assessing an ML product.
02
Predictive modeling & LTV
Build and review machine learning models for demand, churn and customer value. Translate predictions into prioritisation, resource allocation and operational decisions.
Customer value · retention · targeting
03
Marketing mix modeling
Understand which marketing investments create additional sales, where returns diminish and which budget recommendations the evidence supports.
Custom MMM · independent review · budget allocation
04
Causal inference
Establish what a campaign or product change actually causes. Design credible comparisons to measure additional value and separate the effect of your decision from other drivers of demand.
Predictive accuracy alone does not establish the value of an intervention.
TECHNICAL ADVISORY & TEAM DEVELOPMENTA stronger team.
A stronger team.
A clearer roadmap.
I work alongside your team to identify valuable opportunities, choose suitable methods and plan delivery. Recommendations draw on commercial experience: what worked, where it fell short and which risks matter.
Discuss a focused reviewPrioritise the opportunities.
A ranked shortlist based on business value, available data and the team’s capacity to deliver.
Choose the methods and tools.
A suitable modeling approach, clear benchmarks and a practical plan to validate the value.
Build the team’s judgement.
Hands-on project reviews and targeted mentoring, grounded in statistics, with a development plan that helps specialists challenge assumptions and AI-assisted outputs.
02 SELECTED WORK
Experience in practice.
Questions worth asking.
Marketing investment, faster experiments and better customer decisions — drawn from commercial delivery.
MARKETING INVESTMENT · MMM/01
A firmer basis for
marketing investment.
Allocate spend to the activity that creates additional sales. Establish which budget changes the evidence supports before committing more money.
BUDGET DECISION / 01
Should we double the media budget?
Forecast AForecast BPast data
SAME BUDGET. DIFFERENT SALES FORECASTS.+18% to +39%
Forecast growth in ad-driven sales when spend doubles. Both models fit past budgets, but disagree on the return.
OUTCOME
A supported budget rangeclear assumptions and a test plan for uncertain returns
MOBILE GAMING · EXPERIMENTS/02
Faster experiments.
Stronger evidence.
Decide sooner whether a game update increases player revenue. Reduce the number of players needed for a reliable decision, freeing traffic for the next experiment.
EXPERIMENT CAPACITY / 02
Less traffic. A clearer release decision.
−70%potential sample size · same ability to detect uplift
45%narrower confidence interval · if you keep the sample
CountryDevicePre-test revenuePre-test sessions
SMALLER COHORTS OR STRONGER EVIDENCE
Player history removes predictable noise: 70% less estimator variance. Use that gain to recruit fewer players and release experiment capacity, or keep the sample for a more precise decision.
BUSINESS VALUE
−70%potential sample requirement at equivalent statistical power
ADTECH · PREDICTIVE PERSONALISATION/04
Turn relevance
into conversion.
Improve conversion by choosing a more relevant banner for each visitor. Use customer signals, past banner performance and page context to make the decision in real time.
REAL-TIME BANNER SELECTION / 04
Which banner should this user see?
OUTCOME
Real-time personalisationproduction banner selection designed to improve conversion
All case studies draw on my commercial experience. Client details are anonymised and charts use illustrative data.
I’m Jakub Tomaszewski. I specialise in predictive modeling, causal inference and marketing mix modeling, with a strong foundation in mathematics and statistics.
Most of my experience has been in technical leadership roles: setting research direction, selecting modeling approaches and tools, and translating business priorities into data science solutions that teams can deliver and maintain.
Since 2013, I’ve worked across corporations, mid-sized businesses and startups, as well as independently as a freelancer. I remain fully hands-on in model development, technical reviews and mentoring, helping teams build the judgement to connect statistical rigor with commercial value.
INDUSTRY EXPERIENCE
- FMCG
- Mobile gaming
- iGaming
- E-commerce
- Adtech
- Customer service
- Proptech
- Insurtech
Working globally · Contract-based engagementsEnglish & Polish
YOUR NEXT BUSINESS PRIORITY
Put this experience to work.
Bring me a business priority or a technical challenge. We can assess the opportunity, the right approach and what your team needs to deliver.
Discuss your business prioritiesTell me what you are trying to improve and where you need an independent perspective.
DEFINED SCOPE. AGREED TIME BUDGET.
Selected engagements, subject to availability
and conflict checks.