designer

Colleen Marasha — painted designer on the left, photographic engineer on the right

<engineer>

Product thinking and interfaces that feel considered from a magazine cover to a farmer's scan-page.

Offline machine learning, cryptographic systems and payment rails that keep working when the network fails

91.5%
Weighted F1 — Clarion
80.7 KB
Model, distilled from 300 MB
0.837
AUC, up from 0.808
2026
BEng Hons, UZ

A little about me

Beyond the code

I'm a computer engineer who loves numbers which is probably why I'm just as comfortable dealing a hand at the card table. I was raised in a small town, and I've kept its pace with me: I like calm, I like kin, and I have little patience for destructive noise.

Away from the screen you'll find me in the garden coaxing flowers along, on the yoga mat, lost in a novel, or losing gracefully at a board game. I keep cats, quietly and happily, even though my culture side-eyes me for it. And now and then I dance with Elysium Magna because a life of pure logic needs a little rhythm in it too.

Professionally, I work where software meets hardware: machine learning small enough to run on a microcontroller, sensors and actuators in the loop, decisions made on the device rather than in a data centre. My dissertation put a 91.5% F1 audio classifier onto an ESP32-S3 in 80.7 KB, so deaf students get alerted with no cloud and no network at all.

Building for unreliable power and connectivity isn't a limitation I work around — it's the brief I find most interesting. Constraints force honest engineering. Anything that survives them tends to work everywhere else too.

Location:
Harare, Zimbabwe
From:
A small town, at heart
Study:
BEng Computer Engineering, UZ — 2026
Focus:
Edge AI, embedded systems, mechatronics
Loves:
Numbers, novels, flower gardens
Off the clock:
Yoga, board games, a cat or two
Also:
Dancer, Casino Dealer
Colleen Marasha

Portfolio

Some of my latest work

Experience

Where I've worked

Industrial Attachment
Zimbabwe

Software Engineering Attaché

JobSim Enterprise

  • Built the ZIMRA FDMS fiscalisation system: a Flask REST API performing multi-rate VAT reconciliation, cryptographic receipt chaining and Z-report generation against the Fiscal Device Gateway API v6.0 over mutual TLS.
  • Designed and deployed a helpdesk ticketing system that lifted attended customer claims by 95%, analysing usage data to surface recurring issues and improve service delivery.
Dec 2024 – Apr 2025
Remote

Product Management Intern

I_Train Africa

  • Translated survey and market data into learner segmentation insights that directly shaped curriculum and product decisions for digital learning platforms.
2024 – Present
Harare

IT Support Technician

Health Haven Pharmacy

  • Configured branch networks, Wi-Fi and ethernet infrastructure, and advised on hardware procurement — printers, UPS systems, cameras and workstations — for multi-branch expansion.
  • Assessed and provisioned refurbished workstations and a branch server, keeping capital spend low for a growing small business.
Nov 2023 – June 2024
R.G. Mugabe International Airport

Customer Service Agent

National Handling Services

  • Delivered passenger service in a safety-critical airport operation running to fixed departure schedules, resolving disruptions under time pressure.
May – Oct 2023
R.G. Mugabe International Airport

Aircraft Handler — Baggage Operations

National Handling Services

  • Handled aircraft turnaround and baggage operations to strict safety and timing standards.

Toolkit

Languages
PythonTypeScriptJavaScriptPHPSQL
Frameworks
FlaskLaravel 13Next.js 14React
Machine learning
LightGBMXGBoostCatBoostOptunaTensorFlowTFLite
Data & infrastructure
MySQLRedisVercelRailwayCloudflareGitHub Actions
Security
SHA-256RSA-2048ECC ECDSAMutual TLS

Education

NextIn progress

MSc Mechatronics and Artificial Intelligence (MECHART)

University of Zimbabwe

Sept 2026 – present

Postgraduate study in embodied systems: sensors and actuators in the loop, control, and AI running on the machine rather than in a data centre — building directly on the edge-AI work below.

BEng (Hons) Computer Engineering

University of Zimbabwe

Aug 2022 – Aug 2026

Final-year dissertation: Clarion — offline AI alert system for deaf students. Awarded a Distinction. Supervised by Mr G. Jekese.

Certifications

Esports Management
HP / edX · 2024
Game Design
HP / edX · 2024