Lead Product & Experience Designer · AI Engineer · Design Engineer
I work at the intersection of AI engineering and product design — building, evaluating, and shaping AI-powered applications, from retrieval pipelines and LLM experiments to end-to-end product experiences. My work spans healthcare, life sciences, enterprise software, and generative AI products.
With 15+ years in UX and product design, I bring experience leading design alongside engineers, scientists, business analysts, and product teams. My background in mathematics and algorithmic bioinformatics helps me turn complex workflows into understandable, useful experiences.
Portfolio & case studies · LinkedIn · Writing · Devpost
- Agentic UX and human–AI interaction: intent, delegation, approvals, transparency, interruptions, and recovery from errors.
- Applied AI engineering: connecting generative AI APIs with usable mobile applications and product workflows.
- Agentic design: exploring how AI agents can support research, prototyping, evaluation, and continuous improvement.
- Healthcare and life sciences: designing for scientific workflows, complex information, and informed human decisions.
I am expanding my hands-on AI engineering work alongside my 15+ years of product and experience design. I focus on LLM evaluation, retrieval-augmented generation (RAG), information retrieval, reranking, and evidence-based evaluation — connecting model behavior and technical metrics to useful, trustworthy products.
Explore my AI Engineering projects →
| Featured project | What I built | Results |
|---|---|---|
| FilmScope RAG Assistant | Russian-language support assistant with corpus preparation, hybrid BM25 + dense retrieval, cross-encoder reranking, grounded generation, and retrieval evaluation. | Recall@5: 0.9464 · MRR: 0.8296 |
| LLM Product Card Generation | Structured e-commerce card generation with Qwen2.5-1.5B, schema validation, inference benchmarking, concurrency and cost analysis. | 88.33% valid · 5.51 min / 300 · 19.89 RUB/1k estimated |
These case studies document the engineering challenge, implementation, evaluation approach, metrics, and trade-offs.
An AI application that helps merchants create product imagery and videos. I designed, developed, and launched the app, covering UX/UI, React Native development, generative AI integrations, onboarding, and monetization.
Stack: React Native, Expo, Firebase, RevenueCat, and generative AI APIs.
An AI product concept and prototype exploring medication support for individuals, with a longer-term direction in clinical-trial participant monitoring. The design work focuses on understandable medication information, follow-up, and communication between people and care teams.
At EPAM, my work included bioprocess tools, antibody discovery workflows, laboratory interfaces, and enterprise platforms. I connect domain research and product direction with interaction design and engineering collaboration.
A selection of AI and interactive-experience projects, including collaborative hackathon work. Each link includes the original project story, demo, and team credits.
| Project | Short description |
|---|---|
| Bomboloo — a kid-friendly AI assistant | An AI learning prototype that uses animal characters to help children explore weather-appropriate clothing, early learning, and everyday decisions. Built by our team with Gemini, Stable Diffusion, Vue, and Node.js/TypeScript. My role: design, user testing, and project management. |
| VibezAI — MITx Group Project | A collaborative Google AI Hackathon project that turns a simple form into social posts and images for small businesses. Built with Python, Streamlit, Gemini, and Google Cloud Vertex AI to make content creation approachable without advanced prompting skills. |
| mVibe — AI-powered scent platform | A hackathon prototype exploring personalized scent experiences based on music, mood, and context, combining a digital interface with smart-diffuser concepts. The project uses Bolt, OpenAI, Spotify, and Tavus. |
| Cooking Hunt | An interactive Meta Horizon experience where players solve ingredient clues and discover recipes and food cultures. Combines educational game design with mobile/VR interactions using Horizon tools and TypeScript. |
Collaborative research completed in the Bioinformatics Institute's 2018–2019 Algorithmic Bioinformatics program. The institute lists my work under my previous name, Elena Polyakova (now Elena Dalboni da Rocha).
Collaborative bioinformatics research exploring how patterns in DNA promoter sequences relate to gene expression. We developed cisExpress, a computational approach for detecting statistically significant, position-specific regulatory motifs, clustering related sequence patterns, and analyzing their changes over time using Hidden Markov Models. The work also included interactive visualizations to help explore potential regulatory signals relevant to complex diseases.
Researchers: Elena Polyakova (now Elena Dalboni da Rocha) and Daria Balashova. Scientific adviser: Tatiana Tatarinova, University of La Verne.
Project abstract — Bioinformatics Institute (2018/19) (see “Finding of cis-regulatory elements in promoters” under Fall 2018). The historical slide PDF is currently unavailable, so the institute's project record is the primary reference.
A comparative bioinformatics project exploring fish adaptation to deep-water environments through rhodopsin (RH1) sequence analysis. Python and Jupyter notebooks document the exploratory analysis.
Team: Elena Polyakova and Anna Namyatova. Supervisor: Nadezhda Potapova, Faculty of Bioengineering and Bioinformatics, Moscow State University.
Code and analysis: FishGenes · Institute presentation (PDF) · Presentation copy on GitHub
Bioinformatics Institute — graduate profile and project records
These repositories document parts of my development and computational biology background, including educational projects completed during my bioinformatics studies.
| Repository | Focus |
|---|---|
| FishGenes | A collaborative bioinformatics project comparing rhodopsin nucleotide and amino acid sequences in deep-water fish, with Python and Jupyter notebooks. |
| algorithm_bioinf | Algorithmic bioinformatics coursework covering alignments, clustering, and hidden Markov models. |
| ML | Machine learning exercises in Python and Jupyter notebooks, including clustering and model evaluation. |
| NGS | Sequencing-data analysis: read quality, alignment errors, coverage, and De Bruijn graph assembly exercises. |
| scitec_hub | A React/Redux science and technology discovery interface for research, companies, and community content. |
| gravity_doctors | A BigQuery release-notes web application and Google News command-line tool using Python/Flask and Node.js. |
Engineering Has Agentic Engineering. Design Needs Agentic Design.
A perspective on how agents could support the product design lifecycle, with designers defining intent, evaluating evidence, and making decisions.
- Design: Figma, FigJam, product strategy, user research, interaction design, prototyping, and design systems.
- Development: JavaScript, React Native, Expo, Firebase, RevenueCat, and generative AI APIs.
- Computational work: Python, Jupyter, machine learning coursework, and bioinformatics algorithms.
- Education: mathematics and mechanics; postgraduate training in Algorithmic Bioinformatics; MIT xPRO's Designing and Building AI Products and Services program.
Open to Lead/Staff Product Design, AI Product Design, Design Engineering, and applied AI Engineering opportunities, and collaboration on useful AI products.