Building intelligent systems at the edge of neuroscience and finance.

CS @ Johns Hopkins

Baltimore, MD
Currently exploring:Distributed Systems
Hobbies:cinema·flying·dining·safaris·volunteering·nature & outdoors

Right Now

Building

NIH DTI pipeline · EEG-to-image decoder · ML project portfolio

Morgan Anunda

Experience

Industry Experience

Research Assistant

May 2026 – Present

Johns Hopkins School of Medicine · Baltimore, MD

  • Built a 7-stage diffusion-MRI analysis pipeline (Python, Bash) and parallelized it on a Linux HPC cluster with SLURM job arrays, running ~1.9B Monte Carlo streamline simulations as 85 independent jobs, cutting runtime ~30h to <2h (~17x).
  • Traced a false result to a normalization artifact; built a validation layer that recomputed all outputs from source (100% match).
  • Building Spark SQL extraction and quality-control queries on Databricks for a clinical study in a secure research enclave.
PythonBashSLURMDatabricksSpark SQL

Software Engineer

Jun 2025 – Present

AssistiveMath · Nairobi, Kenya

  • Engineering features for a startup backed by a six-figure Mastercard Fellowship and $40K+ in pitching awards.
  • Built an accessible calculator in Vue 3/Nuxt and TypeScript for visually impaired students: expression-evaluation engine, full keyboard control, and MathML output so screen readers speak equations correctly.
  • Fixed WCAG AA contrast and ARIA failures blocking core calculator functions and grew unit tests from 36 to 189 (Vitest), adding Playwright end-to-end and automated accessibility checks.
Vue 3NuxtTypeScriptVitestPlaywrightWCAG AA

Project Manager, Client Consulting Team

May – Jul 2026

diiVe Consulting · Cape Town, South Africa (Remote)

  • Led a 4-person team in a 12-deliverable engagement; delivered final recommendations on time with client sign-off for execution.
  • Conducted financial modeling and led extensive market research to develop a GTM strategy targeting ~$200K USD in ARR by March 2027.
GTM StrategyRevenue ModelingClient ManagementBusiness Discovery

Decoding Subteam Member

Sep 2025 – May 2026

Brain-Computer Interface Society at Johns Hopkins · Baltimore, MD

  • Built an end-to-end PyTorch pipeline to reconstruct images a person perceived from 17-channel EEG; mapped signals to CLIP embedding space to condition Stable Diffusion; set up inference as a Dockerized FastAPI service with CI.
  • Improved reconstruction fidelity 4.8x (CLIP similarity 0.72 vs. 0.15) by evolving the pipeline from a GAN to a CNN encoder and diffusion generator.
PyTorchCLIPStable DiffusionDockerFastAPI

Data & Marketing Analyst

Apr – Jun 2024

Building-U · Canada

  • Managed a $10K/month advertising budget across digital channels.
AnalyticsDigital MarketingPythonData Pipelines

Education

Academic Background

Johns Hopkins University

Aug 2025 – May 2029

B.S. Computer Science · Baltimore, MD

Hodson Trust Scholar, 2025

Activities

National Society of Black Engineers (Senator)Jane Street FTTPCourse Assistant for Gateway Computing: PythonBlack Male InitiativePAVA Center for Entrepreneurship

Projects

Selected Work

optimalLearns Optimal Strategygames playedwins01

MENACE: Tic-Tac-Toe

PythonReinforcement LearningOOPAI History

Donald Michie built MENACE in 1960 using 304 matchboxes and colored beads to prove machines could learn through reinforcement. This project recreates that experiment in Python: a self-teaching agent that starts with zero knowledge and gradually learns optimal strategy through trial and reward. No neural networks, no libraries.

View on GitHub →
MarketTrending02

Hello Finance

AI/NLPPythonFinance API

A trading project that blends classic finance strategies with AI models to track stock trends and read market moods.

10,000trials03

College Monte Carlo

PythonStatisticsData Science

Uses Monte Carlo simulation to quantify expected-value tradeoffs in time-allocation decisions, formalizing intuitive cost-benefit reasoning as probability distributions over outcomes under uncertainty.

View on GitHub →
RISK MONITOR · ACTIVE04

Stroke Prediction

Scikit-LearnMachine LearningPandas

A clean ML pipeline that analyzes health and lifestyle data to predict a patient's risk of having a stroke.

View on GitHub →
NSE EQUITY RESEARCH · NAIROBI SECURITIES EXCHANGEFilingMarketNewsLangGraphpipelineClaudeHaikuBUYHOLDSELLFastAPI · WebSocket · Next.js · 54 tests05

SokoIQ

LangGraphClaude APIFastAPINext.jsWebSocket

Architected a 4-agent LangGraph pipeline synthesizing live NSE market prices, revenue/profit disclosures, and news sentiment into BUY/HOLD/SELL/NEUTRAL investment briefs for 10 NSE-listed equities; used Claude Haiku for thesis generation.

View on GitHub →
HOLC redlining grade vs. predicted food desert rate16.2%A/B Graded25.4%C/D Graded1.57×higher risk72,531 census tracts · Random Forest · SHAP · bias audit06

DesertMap

Random ForestSHAPBias AuditScikit-LearnGeoML

Trained a Random Forest on 72,531 U.S. census tracts to predict USDA food deserts. No redlining data was used, yet tracts historically graded C or D are 1.57x more likely to be flagged — the model learned the geography of 1930s discrimination from present-day features. Four independent SHAP bias audits confirm the signal.

View on GitHub →

Skills

Technical Toolkit

Languages

Python
Java
C++
SQL
Excel
HTML & CSS
TypeScript/JavaScript

Systems & Tools

Git
GitHub
FastAPI
Vue/Nuxt
Linux
Bash
SLURM
Docker
Next.js
Singularity
Databricks
Spark SQL

Leadership

Community & Service

Program Director / Peer Mentor

Jul 2025 – 2026

Education and Social Empowerment Program (EaSEP Kenya) · Kenya

Selected as one of 10 high-achieving Kenyan students for intensive academic and leadership training. Directing program operations and mentoring peers through academic and career development pathways.

Mentor

May 2026 – Present

Yale Young African Scholars

Mentoring students across Africa through college applications, gap-year planning, and connecting to academic and career opportunities.

Advisory Board Member

Aug 2024 – Present

For a Smile's Sake · Kenya

Advise youth-empowerment initiatives (2K+ beneficiaries) and guide a volunteer network across seven regions across Kenya, Uganda and the Diaspora.

Let's Build Something Together.

Say Hello

© 2026 Morgan Anunda · Baltimore, MD

39.3299° N, 76.6205° W