Anastasia Kalacheva

Achievements: • Implemented new CRM with 60K+ records migrated loss-free, doubling operator calls and cutting reporting from 2h to 2min via automated KPI dashboards and multi-channel comms. • Reduced late cancellations by 85% and planning time by 30% through WMS automation, eliminating manual refunds and cutting reconciliation from 2h to 5min. • For Nestlé, designed KPI Accuracy reporting with scoring, filters and automated monitoring, substantially boosting overall traceability and accountability.

Project – Telemedicine Application (mobile & web) Achievements: • Led launch of telemedicine app (billing & scheduling), boosting digital positioning and driving 3.4% YoY revenue growth. • Integrated ERP and enriched 100K+ patient records, cutting workflow time by 25% and improving compliance. • Enhanced patient experience by unifying online/offline data. • Built KPI-based financial model to forecast profitability and scaling. • Enabled smooth adoption by 100+ doctors via structured training sessions.

Project – Electronic Medical Records (EHR) for nationwide U.S. medical billing: • Reduced manual steps in insurance eligibility checks by ~50% via workflow redesign & EHR integration • Cut average registration/verification time by refining patient info flow and simplifying intake UI • Increased satisfaction, reduced errors by turning staff/doctors’ feedback into user stories, validated with A/B tests • Strengthened retention and competitiveness of iSmart EHR with smoother onboarding & digital positioning

Project – B2B Ordering Application for a HoReCa marketplace platform: • Achieved 100% client migration from phone orders to the online B2B platform by conducting interviews and defining critical features (catalog, bulk ordering, recurring templates). • Increased client engagement and order frequency by launching a loyalty program with cashback incentives. • Delivered the B2B ordering app on time and within budget by coordinating a cross-functional team and managing requirements.

• Researched mineral mapping from hyperspectral imagery (hundreds of bands), designing efficient architectures. • Designed and implemented four deep learning approaches: 1D CNN, RNN, semi-supervised learning (SSL), and deep active learning (DAL). • Achieved >98% accuracy with RNN models while optimizing inference time and cost. • Co-authored a forthcoming GFZ paper on hyperspectral data analysis, under submission to an ISI journal.