Liudmila Davydiuk
Senior Data & Research Analyst | QA · AI Output Validation
New to Contra
Liudmila is building their profile!
Founded and operated a fully functional e-commerce apparel project from the ground up — not as a developer, but as the owner and primary stakeholder overseeing every stage of platform development in detail.
Worked closely with the technical team to define requirements, review progress, make build decisions, and ensure the final product matched the project vision. Managed supplier relationships across Europe, handled operational logistics, and owned the full commercial strategy from launch through day-to-day operations.
The experience of being on the stakeholder side of a full-scale technical build — knowing what questions to ask, what to push back on, and what "done" actually looks like — directly informs how I work with technical teams and data systems today.
1
3
Competitive intelligence is only as useful as the decisions it enables. This series of research projects went beyond surface-level competitor profiling — digging into market positioning, pricing dynamics, differentiators, and emerging trends across a defined competitive landscape.
Raw data came from multiple primary and secondary sources. The real work was in the synthesis: cutting through noise, identifying what actually mattered strategically, and structuring findings in a way that senior leadership could absorb quickly and act on confidently.
Deliverables were built for the room — executive-ready reports and presentations that didn't bury the insight in data, but led with it. Each round of research informed real strategic decisions, not just quarterly slide decks.
1
6
Built a user-facing quote and recommendation engine for a mobile insurance app — the kind of tool where accuracy isn't optional and the user experience lives or dies on how fast and clearly the output lands.
The model took structured user inputs and translated them in real time into personalized pricing and plan recommendations. The challenge wasn't just the logic — it was making sure the outputs were consistent, explainable, and trustworthy enough for users to act on without a second thought.
Pricing logic was built to handle edge cases cleanly, with validation checks ensuring no combination of inputs could produce a misleading or erroneous quote before it reached the user.
The end result was an app users found exceptionally intuitive — complex insurance logic made invisible by a clean, frictionless experience that guided users to the right plan without confusion or overwhelm.
0
9
Designed and built a full data pipeline — pulling from multiple sources, resolving inconsistencies, enriching with external variables, and delivering clean outputs through interactive Power BI dashboards stakeholders can actually act on.
Designed and built a full data pipeline — pulling from multiple sources, resolving inconsistencies, enriching with external variables, and delivering clean outputs through interactive Power BI dashboards stakeholders can actually act on.
The starting point was a large, structurally complex dataset with inconsistent formats, missing values, and misaligned fields. The goal was a repeatable pipeline that transformed it into something reliable enough to drive real decisions.
The enrichment layer was the differentiator: layering in demographic and contextual variables from external sources to add depth that the raw data alone couldn't provide. The final output was a Power BI dashboard built for non-technical audiences — clear enough to act on, rigorous enough to hold up to scrutiny.
AI-assisted validation (Claude Enterprise) was used throughout to cross-check outputs and flag inconsistencies before delivery
1
13