MM.

Margarita Makeeva · Scientific software developer

I build software
for discovery.

I turn complex scientific data and research questions into reliable, traceable software—from Python pipelines to interactive applications.

Explore research software GitHub

Philadelphia, PAOpen to relocationSeeking scientific software roles

01observeraw data
02modelfind signal
03understandusable insight
current focusscience
× software

research questionworking application

FocusResearch software
Core toolsPython · FastAPI · React
MethodsValidation · ML · Visualization
Download résumé

Selected research + product work

From messy inputs to thoughtful, working systems.

I care about what happens between the original observation and the final interface: the structure, the assumptions, and the path back to evidence.

01Independent research software · 2026—present

BodyState Mapper

Making uncertain biological evidence traceable.

A Python framework for structured inference across incomplete physiological, biochemical, and neurobehavioral observations. It keeps evidence, uncertainty, and source provenance visible from input to review.

  • Evidence graph + source registry
  • JSON schemas + validation scripts
  • Reproducible review outputs

Python · JSON · CSV · Markdown · pytest

Explore the repository
02Biomedical data research prototype

NeuroBridge-S4

Finding interpretable patterns in small-N longitudinal data.

A graph-learning research prototype that represents biological adaptation as trajectories rather than isolated measurements, with explainable attribution and reference-calibrated interpretation.

  • Graph construction + trajectory analysis
  • Streamlit exploration dashboard
  • Experimental PyTorch autoencoder

Python · PyTorch · Streamlit · Plotly

Explore the research prototype Related Artemis II methodology challenge package ↗
03Scientific data platform

Scientific Data API

One reliable path from heterogeneous files to usable data.

A FastAPI platform for ingesting, cleaning, validating, querying, and visualizing scientific datasets across common research formats.

  • Multi-format ingestion
  • Validation + query-oriented backend
  • Research-ready API architecture

FastAPI · Python · REST · SQL

View the data platform
04Applied product engineering

Truck Repair Assistant

Turning complicated diagnostics into a usable field workflow.

A full-stack application for commercial-truck diagnostics and roadside support, spanning onboarding, protected user flows, service discovery, structured reports, and AI-assisted guidance.

  • End-to-end product workflow
  • Authentication + subscriptions
  • Multilingual diagnostic experience

React · TypeScript · Supabase · APIs

Visit the live product

What I build

Between the lab, the dataset, and the interface.

01

Scientific data systems

Ingestion, cleaning, schema design, validation, and reproducible pipelines for research data.

02

Research APIs

Python backends and clear interfaces that make complex datasets easier to query and use.

03

Explainable ML

Graph-based and uncertainty-aware workflows designed for inspection, not black-box claims.

04

Interactive applications

Thoughtful React and Streamlit tools that bring scientific workflows closer to their users.

Engineering approach

Reproducibility+Source traceability+Validation+Clear human use

About

Scientific curiosity,
software discipline.

I'm a developer with a background in Applied Informatics, working at the intersection of data systems, machine learning, and interactive software.

I'm especially interested in research environments where reliable software helps interdisciplinary teams work with incomplete, heterogeneous, or high-dimensional data.

My web and product background means I think beyond the model or pipeline: I care about documentation, review, usability, and the person who needs to trust the result.

Applied InformaticsHigh-Performance Computing trackEnglish + RussianAuthorized to work in the U.S.