Eric Sekyere Full-stack engineer

I build systems.
You can test them.

Working interfaces, failure modes, and decisions you can inspect. Change an input, follow a request, or challenge an assumption.

Explore the labs Use the terminal

Start here: can extra capacity fix a slow dependency?

Try an engineering decision

Login is slow. What would you try?

A shop’s login service is overloaded. Requests pile up before checkout. Pick a fix and see how many get through.

1. Choose a fix

Compare the same 60-second incident with and without your fix.

Simulation settings

The same seed repeats the same traffic. Leave it at 42 for your first try.

How the simulation works

This uses the Microservice Meltdown engine. “Login” is its Auth service; adding a server adds one replica. Both runs receive 60 requests per second for 60 simulated seconds. Auth slows by 50 ms at second 10; your fix starts at second 12. Request limiting ends at second 40.

Every request is completed, failed, or still queued. This is a repeatable model with fixed routing and immediate scaling, not a production benchmark or a financial cost model.

Investigate in the full lab

Selected projects

Open the app. Inspect the decisions.

All projects

Developer tooling / Personal project

Polyglot Code Sandbox

Run code in four languages without setting up a local development environment.

The challenge
Different runtimes need one consistent editor, error format, and test experience.
The decision
Separate execution from the interface. Use sandboxed frames for browser code, Pyodide for Python, and a remote API for Rust.
What to inspect
Run a failing test or an infinite loop. Structured errors and watchdog timeouts keep failures contained. Rust execution requires the external service.
Run the sandbox Read engineering notes

Systems engineering / Personal project

Microservice Meltdown

Investigate a failing checkout system and measure whether your intervention actually helped.

The challenge
A recovery needs evidence: where requests failed, how queues grew, and what changed after an intervention.
The decision
A seeded simulation with immutable state transitions and an action log makes incidents repeatable and decisions replayable.
What to inspect
Pause, scale a service, then compare with the same-seed baseline. Every request is delivered, failed, or queued. This is a model, with sampled latency metrics.
Investigate an incident

journey.trace

From scripts to systems.

A stack trace of how my engineering work grew. Open a frame to follow the context.

  1. share.tools()Make useful work reusableFeb 2016 – Jun 2019
    • AngularJS
    • ExpressJS
    • APIs

    Built a centralized scripting repository adopted company-wide, and co-developed a tool for testing and validating product concepts.

    The work centered on turning scattered scripts into tools other people could find and use. Shared access, consistent interfaces, and reuse were the engineering concerns.

  2. modernize.systems()Keep existing systems movingJun 2019 – Jun 2021
    • Python
    • Nuxt
    • LibreNMS

    Led the Python 2.7 to 3.11 migration across 50+ applications, re-platformed a B2B site, and built an alert transport for network incidents.

    This work crossed application code and operations: changing old foundations, restoring testing standards, and making faults actionable.

    Explore the alerting lab
  3. ship.web()Connect the interface to deliveryJun 2021 – Jun 2023
    • Vue / Nuxt
    • REST APIs
    • Testing

    Implemented server rendering for a B2B website, integrated APIs with input validation, and established end-to-end tests for a residential platform redesign.

    Rendering, validation, and release testing met in the same product. The work included defining MVP scope with cross-functional teams and testing the paths customers actually used.

  4. connect.products()Make separate systems work togetherJun 2023 – Present
    • React
    • Python
    • ServiceNow

    Built integrations for ServiceNow, DocuSign, and Google Calendar, plus reusable React components backed by Wagtail CMS.

    The scope extended beyond an individual feature: shared components, API listeners, architecture research, and mentoring engineers working across the stack.

The work continues in the experiments on this site. Explore the labs

Interactive portfolio

A different way to explore my work.

Ask about a project, browse the virtual filesystem, or inspect live site health. Commands and questions share one interface.

eric@portfolio~/portfolio
connectederic@portfolio
Eric Sekyere · Senior Software EngineerType any command or just ask a question — the terminal figures it out Tap a chip below, or type help to see everything available.
AI assistant NEW Just type naturally. Commands run here; questions go to AI.
or try a shell command
Tabcomplete·Ctrl+Rsearch·Ctrl+Ppalette

Built with hybrid search over my projects and writing, plus tools for live queries. Read the engineering notes

Engineering labs

Put the systems under pressure.

Repeatable incidents, observable behavior, and evidence you can take away. Explore the models and their tradeoffs.

ENGINEERING LAB / 01

Microservice Meltdown

An incident is unfolding. Find the bottleneck, protect checkout, and measure whether your intervention helped.

  • Request tracing
  • Deterministic replay
  • Baseline comparison
Take the on-call shift ENGINEERING LAB / 02

Terminal Escape

Investigate a compromised Unix box. Follow the evidence through logs and release artifacts, then contain the exposure.

  • Unix permissions
  • JSON pipelines
  • Incident reports
Open an investigation
Explore systems internals Try the AI brand generator Visit the arcade

Let's work together

Have a system to improve
or a product to build?

Let's talk through the architecture, a tradeoff, or what you're building.