I'm Asmaa Abdul-Amin, a software engineer and creator of BLKPVNTHR.OS — decision intelligence infrastructure that connects data, evidence, analytical engines and workflows into systems designed to support traceable, data-backed decisions.
Opens in guest mode — no sign-up.
root@blkpvnthr-os:~$ — and typing
start opens the AI Brain.
featured product
BLKPVNTHR.OS connects data, evidence, analytical engines, workflows and human review into one graph-based system for making traceable, data-backed decisions. Every data source, engine, workflow and experiment is a connected resource, and the architecture is designed so a conclusion can be linked back toward the evidence behind it. The current live implementation includes finance-oriented engines and workflows.
“The graph is the application — every node is clickable, opens the Inspector.”
Six capabilities behind the graph — the decision lifecycle, from connected data to reviewable outcomes. Scroll, swipe, or use the arrows.
Every engine, agent, workflow and data source is a node on one live graph. Selecting a node opens the Inspector, so you navigate the system by exploring the graph rather than hunting through menus.
/osFragmented data, documents and context are pulled into one connected model. A vault holds documents and notes; a Graphify pipeline turns arbitrary input into a knowledge graph the whole system can query.
/vault, /graphifyThe architecture is designed so conclusions can be linked to their supporting data, experiments, assumptions and processing steps — so a result can be followed back toward its source and the reasoning behind it.
evidence, provenanceDomain-specific engines run analysis and scenario evaluation over connected, read-only data, comparing possible actions. Each engine can define its own data sources, evidence requirements and review boundaries without changing the underlying graph-based architecture.
/enginesWorkflows and missions chain steps into repeatable operations, so recurring analysis and handoffs are assembled from parts that already exist rather than rebuilt each time.
/workflowsDecisions, their evidence and the steps behind them are recorded and reviewable. A memory layer preserves context across runs, and consequential actions are designed to route through human review rather than fire on their own.
/memoryTrading is paper-first — live orders stay disabled until a strategy clears the platform’s paper-trading threshold. Bank connections are read-only; the app never sees your bank credentials. Nothing here is investment advice.
selected work
Systems I designed and built, from decision intelligence infrastructure to public-facing science-mission web work.
Four systems. Scroll, swipe, or use the arrows.
Flagship product
Decision intelligence infrastructure: a knowledge-graph interface over connected data, analytical engines, evidence, workflows and human review. The current live build includes finance-oriented modules — a research pipeline, a read-only accounting agent, and a paper-first trading engine.
Space and mission
Public-facing science-mission web and visualization work — mission portals, interactive explainers and outreach visuals. Detail here is limited to publicly released material.
Research platform
A FastAPI research engine with pattern discovery, alpha discovery, evidence fusion and a strategy lab, wired to a paper-first execution surface. Figures shown are demo state, not a track record.
Data and AI
Data ingestion and transformation: a Python project that scrapes retail prices through the SerpAPI search API, with a web-scraping fallback, then uses pandas to compute year-over-year change and plot it. The same pattern — pull messy third-party data, normalize it, make it legible — runs throughout my work.
services
I help organizations design and implement systems for operational intelligence, research automation, knowledge management, data integration, financial reporting, and AI-assisted workflows.
Agent, retrieval and workflow design — where the model belongs, where it does not, and how state moves between them.
Metrics, targets and reporting surfaces built on read-only integrations, designed so numbers can be traced to their source — financial or otherwise.
Ingestion and transformation pipelines, third-party APIs behind server-side functions, secrets kept off the client.
Document vaults, knowledge graphs and memory layers your team and your agents can both query.
Hypothesis-to-validation pipelines with holdouts, robustness checks and a written record of what was rejected and why.
A hard look at an existing system: failure modes, data handling, test coverage, deploy path and what breaks first under load.
React and TypeScript on the front, Python or Node services behind it, deployed and maintained.
The recurring manual work — reports, reconciliations, handoffs — turned into scheduled, observable jobs.
We map the work as it actually happens today, name the bottleneck, and agree on what success would look like.
A narrow, working slice of the system — enough to prove the approach and enough to kill it early if it does not hold up.
The full build: tested, documented, deployed, with the integrations and access controls it needs.
Monitoring, iteration and the next round of scope, once the system has met reality.
experience
Engineering roles across space-mission web infrastructure, undergraduate research, and independent work.
The Johns Hopkins University Applied Physics Laboratory — SES/SOF-2
Fortified and optimized Space Sector web infrastructure across 15+ NASA mission platforms. Engineered data ingestion and transformation pipelines for high-frequency environments.
The Johns Hopkins University Applied Physics Laboratory — REDD/R1N
Engineered and deployed web pages for a NASA Dragonfly mission portal. Built an ML model relating COVID-19 policy measures to mortality data.
Developed the core skills behind my software development and data engineering work, and continue to extend them through self-directed study in machine learning, quantum computing and DevOps.
University of Maryland Global Campus
Comprehensive coursework in data science and engineering, applied to building dynamic and interactive web applications.
2022–presentFoundational computer science, which is where the software development habits started.
2026about
Long before I wrote code I was building things in whatever medium was in reach. I learned piano, guitar and bass entirely by ear — listening, experimenting, recreating what I heard without sheet music — and when I wasn't playing I was painting and drawing, preoccupied with composition, balance, and the work of turning an idea into something you can actually hold. That is where I learned to see patterns and to look for structure inside complexity.
Software turned out to be the same instinct with different instruments. Late-night programming experiments became a real interest in designing systems, and today my work spans backend development, distributed systems, machine learning and data-driven software. That path included an internship at the Johns Hopkins Applied Physics Laboratory supporting space-mission web work. The canvas changed; the mindset didn't. I'm a data-driven engineer, cat person, artist, musician, quantum-curious tinkerer, and quietly ambitious programmer — and creativity is still the engine behind everything I build.
contact
Tell me about the system you want to build, or the one that is already giving you trouble.
Thanks — your message is on its way. I'll get back to you soon.
Something went wrong sending that. Email me directly at blkpvnthr@asmaa.dev.