$ whoami
> Ngana Noa Junior Frederic ("MasterSilver")
> Backend Engineer, Machine Learning Researcher
> Based in Yaounde, Cameroon | Open to relocation an Remote
> Focus: distributed backend systems, applied NLP, mobile platforms
> Also building: algorithmic trading systems, application security research
Backend engineer and machine learning researcher with 2+ years building production systems across healthcare, fintech, and real-time platforms. My core work is backend architecture and API design, and my research work is applied NLP, primarily low-resource language modelling. I have shipped end to end: REST APIs, mobile builds in Expo, and deployment on AWS EC2 and DigitalOcean.
Selected results:
- Built the backend and NLP pipeline for a clinical triage system reaching 93% answer accuracy, backed by 583+ documents indexed in Pinecone with sub-2-second retrieval under load.
- Reduced API latency by 42% on a real-time booking platform through query optimization, indexing, and a caching layer.
- Co-authored NLP research on low-resource African languages (Igbo, Bulu, Hausa, Swahili, and six others), outperforming NLLB-200 and OPUS-MT baselines; published in Informatics (MDPI).
- Built the payments microservice for a cross-border money-transfer platform, reducing failed transactions by 28% through fraud-detection rules.
Outside of client work, I build algorithmic trading systems in Python and MQL5, and study application security through structured penetration testing on my own projects.
Languages
Backend
Machine Learning & NLP
Frontend & Mobile
Data & Cloud
Security & Quant Tooling
Backend Engineer (Lead Contributor), AFreeServ — Sep 2025 to Nov 2025
- Built a real-time messaging system for a service-booking platform serving thousands of concurrent users
- Reduced API latency by 42% through query rewrites, TypeORM indexing, and a caching layer
- Brought test coverage to approximately 85% using Jest and Supertest ahead of a major production release
- Refactored core backend modules, tripling throughput
Backend Engineer, Ukuqala Platform (Clinical Triage System) — Oct 2024 to Present
- Designed the backend for a clinical triage platform that analyzes patient symptoms and surfaces structured guidance for clinicians
- Trained and integrated NLP models mapping free-text symptoms to candidate diagnoses
- Built the mobile application in Expo (React Native), covering the patient intake flow and clinician dashboard
- Indexed 583+ medical reference documents in Pinecone, keeping retrieval under 2 seconds under load
Python Backend Engineer, Primus Cloud Solutions Ltd (Sendi) — Aug 2024 to May 2025
- Built the core transaction backend and cross-border compliance logic for an international money-transfer platform
- Wrote the Stripe payment microservice from scratch, adding fraud-detection rules that reduced failed transactions by 28%
- Built a notifications service (email, SMS, push) with sub-1-second delivery latency
- Deployed a horizontally scalable stack (FastAPI, Redis, PostgreSQL) on AWS EC2 and DigitalOcean
Research Assistant, NLP and Data Science, University of Hagen & ICT University — Jul 2024 to Feb 2025
- Built LinguoMT, a neural machine translation pipeline fine-tuned on MarianMT across 9 African language pairs, including Igbo, Hausa, Swahili, Ghomala, Bulu, and Fulfulde Adamawa
- Achieved a BLEU score of 40.67 on English-to-Igbo translation, outperforming NLLB-200 and OPUS-MT baselines
- Established the first publicly available BLEU benchmarks for several Central African languages
- Released datasets, tokenizers, and fine-tuned models on Hugging Face Hub; co-authored the resulting paper in Informatics (MDPI)
| Project | Description |
|---|---|
| AlgoViz | Full-stack interview-prep platform (Next.js) covering 25 DSA problems with step-through algorithm animations and multi-language code examples |
| LinguoMT | Neural machine translation research pipeline for 9 low-resource African language pairs, released on Hugging Face Hub |
| Algorithmic Trading Engine | Multi-layer confluence trading systems (Python and MQL5) with ATR-based position sizing and strict capital-preservation rules |
| Application Security Lab | Structured penetration testing practice on self-built applications, covering XSS, SQL injection, and IDOR, using Burp Suite and sqlmap |
