Maria Tsfasman (Masha)

AI Research Engineer at Lovemind AI

About

I'm a data scientist and AI researcher by day and a ceramics artist and teacher by night. I help businesses with everything from LLM research & building ML pipelines to optimising internal business workflows with AI tools.

My PhD work at TU Delft taught me how to ask the right questions, design solid experiments, collect and curate complex datasets, and mentor junior researchers. My most recent Data & AI consulting projects (e.g. prototyping AI features for an edtech startup) have shown me how to make an impact with simple data-driven solutions.

I get my energy from human-centred data and working with people from different professional and social backgrounds (brainstorming, pitching, guiding, collaborating). I am looking for a long-term commitment, now that I am settled in the UK.

Experience

  • AI Research EngineerLovemind AI
    Jan 2026 – present
    • Developing vision-language model frameworks integrating AI capabilities with psychological theory and personality profiling.
    • Designing psychometric prompting methods for LLMs, conditioning model outputs to improve performance.
    • Building research pipelines processing conversation transcripts, personality assessments, and behavioural ratings to analyse social perception patterns.
  • AI & UX Data AnalystGoodUp
    Sep – Oct 2025
    • Transformed CRM data transfer and enrichment workflows with Gen AI tools, now in use for finding and connecting to 5000+ marketing leads per quarter.
    • Investigated company workflows, presenting a comprehensive analysis of bottlenecks, potential Gen AI solutions & their impact on the business operations. My solutions are now in use, decreasing team workload by 25%, leaving more room for customer-centred company time.
    • Improved customer-facing data dashboards in Looker (SQL), for clear platform impact assessment.
  • Data & AI ConsultantSignOnSpot
    Jan – Jul 2025
    • Prototyped and implemented an explainable multimodal AI feedback feature using Time-Series Analysis and LLM techniques.
    • Drafted and maintained GDPR-compliant user licence agreements & data privacy statements.
  • AI & HCI Researcher (PhD)TU Delft & Hybrid Intelligence Centre
    Sep 2020 – Dec 2024
    • Orchestrated the creation of a 700GB multi-modal time-series dataset to enable next-generation automatic meeting support.
    • Performed comprehensive quantitative and qualitative statistical data analysis and visualisation (Python, R, SQL).
    • Developed and deployed end-to-end Statistical Modelling (ML) pipelines using Python libraries like NumPy, PyTorch, and TensorFlow on an AI cluster, achieving robust, above-chance performance across 10k+ simulations.
    • Presented research, wrote research reports, mentored junior colleagues in AI, data analysis, and ethical research practices.

Skills

Programming & ToolsPython, SQL, R, MATLAB, Jupyter, Git, Google Cloud, Latex, Databricks, Excel (VBA)

Data ScienceStatistics, Machine Learning, NumPy, Pandas, SciPy, Scikit-Learn, MatPlotLib, statsmodels, PyTorch, Causal inference, Containerization, Parallel computing, Cloud computing (Google Cloud Platform), Feature engineering & selection, Data mining, collection, visualisation & curation, Data ethics (GDPR, AI Act)

LanguagesEnglish (fluent), Russian (C2), French (B2), Italian (B2)

Education

  • Ph.D. Computer Science (AI & HCI)Delft University of Technology, Netherlands · 2020–2025
  • M.Sc. Artificial Intelligence, with DistinctionRadboud University, Netherlands · 2018–2020
  • B.Sc. Computational Linguistics (NLP)National Research Institute HSE, Moscow · 2014–2018

Awards

  • TAILOR connectivity grant awardeeUniversity of Cambridge · 2023
  • Outstanding Early Career Scientist PaperIEEE Ro-Man · 2022
  • Disciplinary Honours awardee, IRCNUTokyo / Radboud University · 2020