// Applied AI/ML Engineer
I'm Anthony.
I build AI systems that work in production.
Open to remote · GMT-4 · Trinidad
About
// Highlights
Five+ years building production ML systems, from edge-deployed computer vision that reads vital signs from a camera feed to agentic RAG pipelines used in real clinical workflows. I sit at the intersection of research, engineering, and production systems.
5+
Years in production ML
1
Granted US patent
2
IEEE Published papers
2
HIPAA compliant systems
PythonPyTorchLangGraphAWS LambdavLLMOpenCVFastAPIDockerQdrant
// Credentials
US Patent 12,525,360Granted Jan 2026
Systems and methods for assessing health measurements from facial video data using rPPG signal extraction and machine learning.
IEEE Published · DS-RAG FrameworkFeb 2026
Proposed a decentralized RAG protocol enabling Small Island Developing States to maintain data sovereignty while leveraging state-of-the-art LLMs.
IEEE Published · AI Personalization StrategyFeb 2026
Analyzed AI-driven personalization as a blue-ocean strategy in LATAM healthcare and education markets, with focus on agentic AI and Digital Twins.
// The Pipeline
I take problems from first principles to production and own every step in between.
Research
I turn problems into IP. Co-inventor on a granted US Patent in computer vision and author of 2 IEEE-published papers on LLM systems.
PatentsIEEE papers
Engineer
I turn messy data into working models. Five years delivering HIPAA-compliant AI pipelines, computer vision models, and agentic RAG systems in production.
PipelinesAPIsModels
Deploy
I build systems that ship. Experience ranging from model optimization and automated CI/CD to owning compliant production deployments.
InfraCI/CDOn-call
// Career
Experience
Jump to
~/experience/01_hibiscus.md
Lead AI/ML Engineer
Most recentHibiscus Health
Aug 2023 — Jan 2026
Washington DC · RemoteOwned end-to-end ML architecture for a digital health product. Built a patented CV pipeline extracting rPPG signals from facial video and agentic RAG pipelines for clinical nutrition guidance, deployed via AWS Lambda to a live mobile app.
- 8×Inference speedup
- 0.8+AUC health models
- 50GBVideo processed
- 1US patent
// Focus
Computer vision
Multi-threaded rPPG pipeline predicting hypertension, cholesterol & diabetes
MLOps
HIPAA-compliant ETL, CI/CD via GitHub Actions, AWS Lambda deployment
LLM / RAG
Agentic LLaMA pipelines for nutritional guidance and meal planning
Architecture
Designed full system across product and engineering, authored investor docs
// Stack
- PyTorch
- OpenCV
- LangGraph
- vLLM
- Qdrant
- AWS Lambda
- LangFuse
- HIPAA ETL
- GitHub Actions
~/experience/02_precise.md
AI Engineer
Precise Behavioral
Nov 2022 — Aug 2023
Westlake Village CA · RemoteFine-tuned a BERT-based classifier on 5,000+ clinical journal entries to predict anxiety and depression probabilities. Built a HIPAA-compliant FHIR ingestion pipeline and a production drift monitoring system that auto-flagged anomalies.
- 93%F1 score
- -20%False positives
- 5k+Training entries
- HIPAACompliant
// Focus
NLP classifier
BERT fine-tuned on clinical text, evaluated with the Checklist Framework
FHIR pipeline
Python ETL for ingesting EHR records with full HIPAA compliance
Drift monitoring
Lightweight system auto-flagging model anomalies in production
Evaluation
Reduced false positives by 20%, improved vocabulary coverage to 93% F1
// Stack
- BERT
- Transformers
- FHIR
- Python
- HIPAA ETL
- Checklist Framework
- Drift monitoring
~/experience/03_rose.md
Junior ML Engineer
Intern → engineerRose: Smarter Mental Health (Acquired)
Jun 2018 — Nov 2022
New York NY · RemoteBuilt a content recommendation system serving 200+ articles and blogs from user behavioral signals, quantized an ONNX language model cutting inference latency by 25%, and integrated PyTorch NLP models into a Django/Celery backend. Joined as a part-time Data Science Intern during undergrad in 2018; promoted to Junior ML Engineer on graduating in 2021.
- -25%Inference latency
- 200+Articles served
- 4yrTenure
// Focus
Full stack
React/TypeScript frontend with Redux Saga, Django REST backend
Model optimization
Quantized ONNX language model, cutting inference latency by 25%
Recommendation
Content system serving 200+ articles & blogs from user behavioral signals
NLP integration
PyTorch NLP models served async via Celery task scheduling
// Stack
- Django
- React
- TypeScript
- ONNX
- PyTorch
- Celery
- Redux Saga
- REST APIs
// Projects
Selected work.
Real projects, real patents, real papers, real production traffic.
CircuitNet
Turns hand-drawn circuit sketches into digital schematics — OpenCV component detection, custom-CNN classification, then schematic generation.
- 0.96weighted F1
- <400mson CPU
- <15MBmodel
rPPG Vitals
Reads rPPG signals from facial video to predict hypertension, cholesterol, and diabetes risk — deployed in a live mobile app.
- 8×inference speedup
- 0.8+AUC
- 50GBvideo processed
FairWage-ML
An ML pipeline benchmarking global wage fairness against macroeconomic indicators, with counterfactual career pathways generated via DiCE-ML.
- 0.98R²
- 250ksalary records
- 62.7%counterfactual gain
// Academic Journey
Education
MSc Data Science
The University of the West Indies
2025 – 2027 (expected)St. Augustine, Trinidad
- Machine Learning
- Big Data Analytics
- Statistical Inference
- Data Visualization
- Bayesian Methods
- Research Practicum
BSc Electrical & Computer Engineering
The University of the West Indies
2017 — 2021St. Augustine, Trinidad
- Signals & Systems
- Embedded Systems
- Digital Design
- Instrumentation
- Capstone Project
// CONTACT · OPEN TO WORK
Let's build something that ships.
Open to full-time remote AI/ML engineering roles. If you're building something that needs to actually work, let's talk.