Thegretia
Identity & Specialization

Patrick Thomas MBONJO ETIA

Pseudonym: Thegretia • Data Engineer, AI & Machine Learning Specialist, and Python Software Architect.

I am a technical architect and builder operating at the intersection of large-scale data systems, production AI/ML inference, and high-concurrency Python backends.

Over the years, I have architected multi-terabyte Lakehouse pipelines on Azure Databricks, engineered low-latency speech recognition and translation engines for under-represented African dialects, and constructed sub-50ms financial streaming anomaly engines.

My engineering philosophy is rooted in measurable business impact: eliminating cloud compute waste, guaranteeing pipeline data freshness, and deploying AI models that operate reliably without unbounded server infrastructure costs.

Quick Snapshot
Focus:Data Pipelines & AI Engines
Primary Languages:Python, SQL, TypeScript
Core Frameworks:PySpark, PyTorch, FastAPI, Delta Lake
Status:Available for select projects
Technical Matrix

Core Competencies & Tooling

Data Engineering & Lakehouse

  • PySpark & Apache Spark
  • Azure Databricks
  • Delta Lake (Medallion Architecture)
  • dbt-core / dbt-databricks
  • SQL (Advanced Optimization & Tuning)
  • ETL / ELT Pipeline Architecture
  • Azure Data Factory (ADF)
  • Great Expectations & Data Quality

AI & Machine Learning

  • PyTorch & TorchAudio
  • Hugging Face Transformers & PEFT
  • Whisper & Wav2Vec2 (ASR / NLP)
  • ONNX Runtime & INT8 Quantization
  • LLM Orchestration & Embeddings
  • Scikit-learn & XGBoost
  • Real-Time Streaming ML & Anomaly Detection
  • Model Profiling & Latency Optimization

Software & Cloud Infrastructure

  • Python (Asyncio, Pydantic, Modern 3.12+)
  • FastAPI & High-Concurrency APIs
  • Docker & Container Orchestration
  • Redis (Streams, Caching, Lua)
  • Apache Kafka & Streaming Events
  • PostgreSQL & Supabase
  • Linux Server Administration
  • Git & GitHub Actions CI/CD
Track Record

Professional Experience & Milestones

Solutions Architect & Solution Trainer

GEMULA

Apr 2026 — Present

Designing robust technical architectures, leading hands-on technical demonstrations, and delivering comprehensive technical training programs while contributing to high-level strategic decisions.

Architected end-to-end technical solutions aligned with client requirements and strategic business goals.
Delivered live technical demonstrations and solution walkthroughs to key stakeholders.
Developed interactive training materials, practical coding labs, and technical documentation to upskill teams.

Software Engineer & Partner

The Intelligent Eyeglass for the Blind (IEGB)

Jul 2025 — Present

Designing and deploying embedded assistive technology solutions powered by AI, Machine Learning, Cloud Foundations, and Python to empower visually impaired individuals.

Architected and deployed an embedded platform for smart eyeglasses providing real-time sensory assistance to non-sighted users.
Engineered machine learning pipelines and cloud backend integrations using Python and robust cloud foundation services.
Collaborated across multidisciplinary teams to translate accessibility requirements into high-impact, user-centric software features.

Data Engineer (Freelance)

Orange Cameroun

Jun 2026 — Present

Engineering end-to-end data pipelines for multilingual voice AI systems, specializing in the collection, curation, and preprocessing of audio and linguistic datasets for Speech-to-Text, Machine Translation, and Text-to-Speech models.

Designed robust ingestion and ETL pipelines in Python to transform raw, heterogeneous speech recordings into clean, model-ready datasets for local languages.
Curated and structured high-quality audio corpora tailored for training Speech-to-Text (STT) and Text-to-Speech (TTS) neural architectures.
Collaborated within a cross-functional team to integrate local language data into scalable production-grade AI solutions.

Ready to build high-performance data & AI systems?

Whether you need to scale Lakehouse ETL pipelines, deploy custom AI models, or architect resilient Python backends, let's collaborate.

Get in TouchView Case Studies