Hi, I'm Solo Shun.

I'm an AI/ML Engineer

I design and deploy intelligent systems that solve real-world challenges across AI and finance. From building autonomous AI agents and production-scale ML & Data pipelines to developing quantitative trading systems and algorithmic strategies for financial markets. Let's build the future together!

About.

Hey there! I'm Solomon Eshun (but I love to be called Solo Shun 😄). I'm an AI/ML Engineer with a strong Software Engineering foundation, currently working as an DataOps Engineer at EnBW (via Ishango AI, London UK) and Founding ML Engineer at Noeud.

I specialize in building advanced AI systems that solve real-world problems - from autonomous research agents that rank among the top globally, to production-scale DataOps and MLOps pipelines serving enterprise European clients.

At MinoHealth AI Labs, I built the Moremi Deep Research Agent, achieving a RACE score of 46.3 on the DeepResearch Bench (ranked 3/18 agents). I've architected systems processing 20,000+ molecules and protein sequences in multi-day runs, and co-authored 4 research publications in 2025 on AI-driven drug discovery and biosecurity.

My tech stack spans PyTorch, TensorFlow, FastAPI, Docker, and the full DataOps and MLOps toolchain (DataDog, Prefect, MLFlow, CI/CD etc.). I'm also proficient in full-stack development with React, Next.js, Node.js, Prisma, and building scalable distributed systems.

When I'm not training models or debugging pipelines, I'm deeply into the world of quantitative finance and algorithmic trading - building forex trading strategies with MQL5, Python, MetaTrader, and TradingView. I love applying machine learning to financial markets and exploring time-series forecasting.

Outside of AI and fintech, I'm a gamer at heart 🎮 - gaming helps me set my mind free and recharge. I'm deeply passionate about motivating people to bring out the best in themselves, reading self-help books for personal growth, and listening to motivational talks from my mentors. I also love public speaking and seize every opportunity to inspire and share knowledge on stage.

I mentor at bootcamps, contribute to open-source AI education projects, and enjoy watching science and tech videos to stay curious. Let's connect and build something amazing! 🚀

My links

Programming, AI/ML & Data Science

PythonC++JavaPyTorchTensorFlowTransformersLLMsComputer VisionYOLOPandasNumPyScikit-learnTime-SeriesXGBoostLSTMReinforcement LearningAgentic AIRAG Systems

DataOps, MLOps, SRE & DevOps

DockerDataDogPrefectIaaS (AWS, Azure, DigitalOcean)CI/CDMLFlowFastAPICeleryRabbitMQRedisPostgreSQL & MySQLMongoDBGitLinuxETL PipelinesData Modeling & Database Design

Full-Stack, Design & Trading

ReactNext.jsNode.jsTypeScriptJavaScriptTailwindReact QueryFigmaPrismaJustinmindMQL5MetaTraderTradingViewForex TradingAlgorithmic Trading

Projects.

An image of the Unified Bioinformatics Orchestration Infrastructure project.

Unified Bioinformatics Orchestration Infrastructure

Docker - FastAPI - Celery - RabbitMQ - MongoDB - Redis - Python - C++ - Java - Perl

Scalable distributed infrastructure unifying 20+ heterogeneous bioinformatics tools. Powers autonomous agents, web platforms, and internal analytics.
Learn more >

An image of the Moremi Bio Co-Researcher project.

Moremi Bio Co-Researcher

Python - Agentic AI - LLMs - Bioinformatics Tools - Security Hardening

Public-facing bioinformatics agent on moremi.ai, providing secure access to the unified infrastructure for computational biology research.
Learn more >

An image of the Moremi Deep Research Agent project.

Moremi Deep Research Agent

Python - LLMs - Agentic AI - RAG - FastAPI - Research Automation

Advanced autonomous research agent achieving RACE score of 46.3 on DeepResearch Bench (ranked 3/18 globally).
Learn more >

An image of the Moremi Bio Agent project.

Moremi Bio Agent

Python - Agentic AI - LLMs - Prompt Engineering - Tool Use - Chain-of-Thought

Autonomous AI agent leveraging the unified bioinformatics infrastructure for independent drug discovery research. Co-authored 4 publications in 2025.
Learn more >

An image of the Moremi Biokits project.

Moremi Biokits

In Development
Next.js - TypeScript - 3Dmol.js - FastAPI - React Query - MongoDB

Interactive web platform with 3D molecular visualization, providing researchers intuitive access to the unified bioinformatics infrastructure.
Learn more >

An image of the Moremi AI Agents project.

Moremi AI Agents

Python - LLMs - Real-time Search - Medical Imaging - Computer Vision - FastAPI

Suite of general-purpose AI agents on moremi.ai: real-time clinical search and medical image analysis for healthcare workflows.
Learn more >

An image of the Quantitative XAU/USD Session Strategy project.

Quantitative XAU/USD Session Strategy

Python - XGBoost - LSTM - Transformers - Time-Series - MetaTrader5

End-to-end ML pipeline benchmarking XGBoost, LSTM, and Transformer architectures for financial market prediction.
Learn more >

An image of the Building LLM from Scratch project.

Building LLM from Scratch

Python - PyTorch - Transformers - Tokenization - Attention Mechanisms

Educational journey developing a functional LLM from first principles, documented with code and articles.
Learn more >

An image of the OpenDSA - Algorithm Visualization Platform project.

OpenDSA - Algorithm Visualization Platform

Next.js - TypeScript - Tailwind CSS - Framer Motion - D3.js - Canvas API - Monaco Editor - Zustand - Algorithms - Data Structures

A complete modernization of the DS.AlgoDeck project. An interactive visualization platform for understanding algorithms through step-by-step animations.
Learn more >

An image of the DS.AlgoDeck - Interactive Learning Platform project.

DS.AlgoDeck - Interactive Learning Platform

React - JavaScript - Gemini API - Data Structures - Algorithms

Open-source educational platform for learning data structures through interactive visualizations and AI-powered code generation.
Learn more >

An image of the Sonu AI - Plant Disease Detection project.

Sonu AI - Plant Disease Detection

Next.js - ONNX - Computer Vision - PWA - Tailwind - Leaflet

Offline-capable PWA for crop disease detection achieving 96% accuracy, built in 5 days for Ghanaian farmers.
Learn more >

Experience.

Dec 2025 - Present
DataOps Engineer (Data Platform Engineering)
Karlsruhe, Germany (Remote)

Supporting the development of an internal data platform for a leading European energy provider. My work focuses on building shared abstractions and guardrails that standardise data pipeline development across teams. Instead of writing ad-hoc orchestration logic, downstream users rely on platform-managed libraries that handle data quality checks, retries, observability, alerting, and documentation by default. This enables reliable, scalable, and compliant data workflows in regulated, high-availability environments.

MLOpsPythonDockerDataDogPrefectAzure DevOpsCI/CDAWSModel DeploymentSREEnterprise ML
Oct 2025 - Present
Founding Machine Learning Engineer (R&D)
Accra, Ghana (Part-Time Remote)

Leading research and prototyping of real-time financial forecasting systems and deep time-series models. Designing end-to-end pipelines from data ingestion to deployment and continuous retraining for live market prediction systems.

PyTorchTime-Series ForecastingML PipelinesData PipelinesSystem DesignModel DeploymentMLOpsSREPythonQuantitative Finance
Oct 2024 - Dec 2025
Machine Learning Engineer
Accra, Ghana

Engineered unified, containerized infrastructure integrating 20+ bioinformatics and computational biology tools (written in C++, Python, Java and Perl) with asynchronous task orchestration (Celery, RabbitMQ, Redis). Architected Moremi Bio Co-Researcher, Moremi Agent and Moremi Search powering research workflows on moremi.ai platform. Built Moremi Deep Research Agent (RACE score 46.3, ranked 3/18 on DeepResearch Bench) and designed Moremi Bio Agent, an autonomous research agent technology behind 4 co-authored 2025 research publications including a double-blinded validation study with Imperial College London, Infectious Disease.

PythonDockerFastAPICeleryRabbitMQRedisPyTorchLLMsAgentic AIReact3Dmol.jsBioinformatics
Jan 2025 - Dec 2025
AI Engineer & Drone Mapping
Accra, Ghana (Part-Time Contract)

Deployed YOLOv12 computer vision model for cashew stress detection across 42 farms in 28 communities as part of GIZ, MOVE, and FAIR FORWARD funded CADI AI pilot. Led end-to-end drone mapping operations, collecting and annotating 148GB agricultural dataset.

YOLOComputer VisionPythonPyTorchRoboflowDrone MappingAgricultural AI
Jul 2025 - Sep 2025
AI Instructor
Remote (Contract)

Delivered the 2025 Generative AI & Python Summer Camp for students from Montserrat, Grenada, and Dominica. Taught hands-on sessions on Python programming, large language models, vision-language models, and AI agents.

PythonLLMsVision-Language ModelsAI AgentsTeachingCurriculum Development

Publications.

Moremi Bio Agent: Using Neisseria meningitidis Reference Data For The Double Blinded Validation of A General Purpose Biology-Trained Reasoning Model for Pathogen and Antigen Discovery

Co-Authored•bioRxiv•2025
MinoHealth AI Labs • Imperial College London, Infectious Disease, London, UK

First-of-its-kind AI system demonstrating blind antigen inference from antibody sequences. Moremi Bio Nano successfully identified experimentally validated antigens in 3 out of 4 independent tests, correctly predicting SARS-CoV-2 Spike RBD, Neisseria meningitidis fHbp v1.1, and SARS-CoV Spike as top-ranked candidates—marking a breakthrough in reasoning-driven antigen discovery without prior biological context.

DOI: 10.64898/2025.12.17.694980

Moremi Bio Agent: Leveraging Agentic Large Language Model for the Discovery of Broad-Spectrum Antibiotics for Enterobacteriaceae

Co-Authored•bioRxiv•2025
MinoHealth AI Labs

Autonomous AI pipeline for dual-target antibiotic discovery addressing antimicrobial resistance. Generated 1,002 candidate molecules targeting both FabI enzyme and AcrAB–TolC efflux pump simultaneously. Results: 774 compounds passed ADMET benchmarks, 391 exhibited moderate binding to both targets, and 60 top candidates (score ≥ 0.8) showed favorable drug-likeness with minimal toxicity—demonstrating feasibility of AI-driven antibiotic discovery for multidrug-resistant pathogens.

DOI: 10.1101/2025.08.21.671656

Can Large Language Models Design Biological Weapons? Evaluating Moremi Bio

Co-Authored•arXiv•2025
MinoHealth AI Labs

Critical biosecurity assessment revealing dual-use risks of LLMs in biodesign. Generated 1,020 novel toxic proteins and 5,000 toxic small molecules, with all proteins scoring high in toxicity and several closely matching known toxins (ricin, diphtheria toxin, snake venom proteins). Findings challenge claims that LLMs cannot design bioweapons and underscore the urgent need for robust governance and technical safeguards to balance innovation with biosecurity imperatives.

arXiv: 2505.17154

Moremi Bio Agent: A Large Language Model for General-Purpose High-Throughput De Novo Protein and Compound Discovery, Design and Validation (Early Validations)

Co-Authored•Technical Report•2025
MinoHealth AI Labs

A breakthrough multimodal foundation model (mLLM) designed for Artificial General Intelligence for Biology (AGI4Bio). Moremi Bio functions as a general-purpose system for high-throughput, autonomous discovery, design, and in-silico validation of novel proteins and small molecules. By eliminating human-in-the-loop bottlenecks, the agent achieves up to an 11,000x speed improvement over traditional computational methods—reducing discovery timelines from months to minutes for targets including Malaria, HIV, and various cancers.

Moremi Bio Agent: Application of A Foundation Model and End-to-End Automation in the Design and Validation of Monoclonal Antibodies Targeting Plasmodium falciparum Invasion Complex

Co-Authored•bioRxiv•2025
MinoHealth AI Labs • West African Centre for Cell Biology of Infectious Pathogens (WACCBIP), University of Ghana, Accra

AI-driven antibody design pipeline targeting malaria's invasion mechanism. Generated 999 monoclonal antibodies targeting the AMA1-RON2 complex, with 864 showing successful target interactions. Achieved exceptional binding affinities with dissociation constants (Kd) in femtomolar to attomolar range (ΔG: -116.8 to -5.6 kcal/mol), indicating highly stable interactions. Epitope mapping revealed conserved regions for cross-strain efficacy, establishing a scalable framework for therapeutic development against malaria and other infectious diseases.

DOI: 10.1101/2025.02.12.637967

Activities & Certifications.

Conferences & Seminars

Tech & Beyond Expo

Academic City University•2025

Presented on Software Engineering in the Age of AI

Ghana Data Science Summit 2025 (IndabaX)

Ashesi University, Berekuso - Ghana•2025

Guided participants through hands-on LoRA-based fine-tuning and VLM exercises

DS-I Africa Program

University of Ghana, Accra - Ghana•2025

Exhibited Moremi AI from minoHealth AI Labs, a generalist AI doctor democratizing access to quality healthcare across the continent

Volunteer & Mentorship

Global Code 3-Weeks Bootcamp 2025

University of Cape Coast - Ghana•2025

Mentored participants in Python, IoT, and AI integration, supporting capstone projects

Certifications & Awards

GlobalCode Certificate

3-week Programming Camp (Raspberry Pi & Python)•2024

Intensive programming bootcamp focused on hardware programming and Python

Professional Certificate in Robotics, Arduino & IoT

Professional Certification Program•2023

Comprehensive training in embedded systems and IoT development

Best Participant Award

Robotics and IoT TransforME Project•2023

Awarded for exceptional performance and innovation in robotics and IoT project

Contact.

Shoot me an email if you want to connect! You can also find me on LinkedIn or GitHub if that's more your speed.