Computer Engineering student building intelligent systems at the intersection of:
Data → Machine Learning → Computer Vision → Embedded Systems
- 🔬 Final-year project: Contactless Vital Signs Monitoring System — computer vision + rPPG + planned ESP32-S3 wearable
- 🧮 Background: Cybersecurity internship and automation/data work in Python
- 🖥️ Daily drivers: Linux, Python, C++, SQL, Docker, Git
- 📫 Reach me: LinkedIn · anthonyogbuah@gmail.com · Portfolio
How I think about end-to-end intelligent systems — from raw signals to hardware:
flowchart LR
A["📡 Data Acquisition<br/>Sensors · APIs · Datasets"] --> B["🧵 Data Engineering<br/>Pipelines · ETL"]
B --> C["🧮 Processing<br/>Cleaning · Features"]
C --> D["🧠 Models<br/>ML · Deep Learning"]
D --> E["👁️ Computer Vision<br/>Inference · rPPG"]
E --> F["⚙️ Edge Deployment<br/>Quantization · BLE"]
F --> G["🔌 Embedded Hardware<br/>ESP32-S3 · Sensors"]
| 🔬 Computer Vision | 🧠 Deep Learning | 📊 Data Engineering | ⚙️ Embedded Software |
| 📡 IoT / BLE | 🤖 Edge AI | 🔒 Cybersecurity | 🛠️ Linux Systems |
I'm interested in the full path an intelligent system takes: data acquisition → processing → modelling → inference → hardware/edge deployment. My current work spans contactless health sensing with computer vision, data pipelines, and embedded sensor systems.
Final-year engineering project — computer vision + signal processing for non-contact vital signs, with a planned wearable band.
Vision pipeline (implemented):
Webcam → Computer Vision → Face / ROI Detection → Signal Processing → Vital Sign Estimation
Wearable integration (in development):
ESP32-S3 → MAX30102 → MPU6050 → BLE → Monitoring Station
What's implemented: face detection (MediaPipe, 478 landmarks) · heart rate via rPPG (POS + FFT on forehead ROI) · respiration via chest-ROI optical flow · pupil dilation (iris-calibrated + EAR blink) · age estimation (MobileNetV3-Small transfer learning on UTKFace) · PyTorch with ONNX export.
In development: ESP32-S3 wearable band with MAX30102 + MPU6050 sensors over BLE, MLX90614 IR temperature.
Tech: Python · MediaPipe · OpenCV · PyTorch · NumPy · Signal Processing · rPPG · ESP32-S3 · BLE
Status: research / educational prototype — 57 unit tests passing. Not a medical device and not clinically certified.
📊 data-ingest-ETL — Data Engineering
Multi-tenant ingestion & ETL platform: event API, S3 file ingestion with delta loading (SHA-256 fingerprints, schema validation, dedup), watermark-based DB puller, Dagster orchestration, dbt staging → marts, Terraform on AWS, CloudWatch/SNS observability.
Python Dagster dbt PostgreSQL Terraform AWS Docker
🔍 sql-optimizer-cli — Systems / Data Tooling
Unix-native CLI that analyzes SQL against a live database — real schema, indexes, and EXPLAIN plans — returning prioritized recommendations: security issues, index DDL, partitioning, rewrites, workload regression tracking, and a TUI dashboard. Supports PostgreSQL, MySQL, SQLite.
Rust SQL PostgreSQL CLI Static Analysis
🤖 internship_automation — Automation / AI Systems
Autonomous agent that discovers opportunities, applies via web forms (tiered Playwright + LLM fallback), and sends self-checked cold emails. Event-sourced, replayable actions, FastAPI + arq + Postgres/pgvector, Docker deployment.
Python FastAPI Playwright PostgreSQL Redis Docker
🎵 lyricqueue — Deep Learning / Recommender Systems
Music recommendation engine that predicts "sonic" characteristics from lyrics alone via an RNN/Transformer, combined with multimodal embeddings (lyrics + mel-spectrograms + metadata) in a hybrid two-stage pipeline.
Python PyTorch Transformers FAISS FastAPI
🛡️ network-traffic-detection-anomaly — ML / Cybersecurity
Network traffic anomaly detection applying ML algorithms to identify suspicious patterns — where my cybersecurity internship interests meet applied machine learning.
Python ML Network Security
📊 antlyst — Data Analytics / Visualization
Data visualization tool (from “anthony + analyst”): users upload CSV files and get data visualization dashboards in any dashboard style in under a minute. Bridges analytics and ML-style plots in a quick, no-setup workflow.
TypeScript Data Visualization CSV Analytics Dashboards
🧵 Data Engineering
- Languages & Tech Stack Python, SQL, PostgreSQL, AWS Services
- Processing: Pandas, NumPy, ETL/ELT design, delta loading, watermark-based sync
- Orchestration: Dagster jobs and schedules, dbt staging → mart layers
- Ingestion: REST APIs, S3 event-triggered pipelines, file schema validation
- Evidence: data-ingest-ETL
🧠 Deep Learning & AI
- Frameworks: PyTorch (used in health-project and lyricqueue), Scikit-learn
- Models: MobileNetV3-Small transfer learning (UTKFace), RNN/Transformer lyric encoders
- Optimization: ONNX export, INT8 quantization, CPU inference benchmarking
- Techniques: hybrid recommenders, embeddings, attention-based sequence models
👁️ Computer Vision
- Face & landmark: MediaPipe Face Landmarker (478 points), ROI extraction
- Motion: OpenCV Farneback optical flow, EAR-based blink detection
- Biomedical signals: rPPG (POS algorithm + FFT), respiration from chest motion
- Evidence: health-project
⚙️ Embedded Systems & IoT
- Platforms: ESP32 / ESP32-S3, embedded C/C++ (coursework & labs)
- Sensors: MAX30102 (pulse oximetry), MPU6050 (IMU), MLX90614 (IR temperature) — hardware integration planned for the wearable band
- Connectivity: BLE for wearable-to-station sync, IoT data collection
- Focus: hardware/software co-design, real-time constraints, edge inference
- Evidence: health-project (wearable in development), school-labs
🛠️ DevOps / Infrastructure
- Containers & CI: Docker, Docker Compose, GitHub Actions
- IaC: Terraform modules (VPC, RDS, ECS, Lambda, S3) in data-ingest-ETL
- Backend: FastAPI, Redis task queues, PostgreSQL (incl. Neon/pgvector)
- Tooling: Linux, Git, GitHub, Jupyter, VS Code
🔒 Cybersecurity
- Cybersecurity internship — hands-on exposure to security operations
- ML-based network anomaly detection (network-traffic-detection-anomaly)
- Interest in secure-by-design systems, injection prevention, and secrets management
Languages: Python · C/C++ · SQL · Rust · TypeScript
Data: PostgreSQL · Pandas · NumPy · Jupyter · dbt · Kafka
AI / ML: PyTorch · Scikit-learn · OpenCV · MediaPipe
Embedded: ESP32-S3 · BLE · Sensor interfaces (MAX30102, MPU6050) · Embedded C/C++
Infrastructure: Linux · Docker · Git · GitHub · Terraform
| Area | Why |
|---|---|
| Edge AI & quantization | Getting models to run well on constrained hardware |
| Embedded Linux & real-time systems | Reliable firmware and low-latency sensing |
| MLOps | Training-to-deployment pipelines for CV models |
| Distributed data systems | Scaling pipelines beyond a single node |
| Advanced deep learning | Attention architectures, multimodal models |
Last refreshed 2026-10-05 13:30 UTC by GitHub Actions.
- yesterday — pushed 1 commit(s) to
limit-triggersto anthonyy616/weltrade-final-strat-queued - yesterday — pushed 1 commit(s) to
testto anthonyy616/antlyst - yesterday — pushed 1 commit(s) to
limit-triggersto anthonyy616/weltrade-final-strat-queued - yesterday — pushed 1 commit(s) to
limit-triggersto anthonyy616/weltrade-final-strat-queued - yesterday — pushed 1 commit(s) to
limit-triggersto anthonyy616/weltrade-final-strat-queued - yesterday — pushed 1 commit(s) to
mainto anthonyy616/antlyst
