I specialize in building end-to-end AI systems, from serverless data pipelines to production-grade ML models. My work focuses on automating complex workflows using Google Cloud Platform (GCP), Vertex AI, and Large Language Models.
An AI-powered financial analysis platform that generates actionable trading signals for the Russell 1000. It automates the ingestion of financial data, analyzes it using custom ML models, and surfaces structured Call/Put setups.
The Code Behind the Platform:
- profitscout-engine: The serverless backend built on GCP. Handles data ingestion, pipeline orchestration, and signal generation.
- profitscout-models: The ML pipeline using Vertex AI for feature engineering and training.
- profitscout-gpt: FastAPI backend serving AI-driven financial research and summaries.
RAG & Vector Search A research paper recommendation engine powered by arXiv.
- Tech: Gemini Embeddings, Vertex AI Vector Search, Cloud Run.
- Features: Harvests papers from CS domains (AI, CV, RO), embeds abstracts, and serves semantic search results via API.
Computer Vision in Healthcare A dermatology second-opinion MVP designed to assist clinicians.
- Tech: Vertex AI AutoML Vision, SHAP (Explainable AI), Firebase.
- Features: Triage prototype for dermoscopic images with explainability overlays.
Computer Vision Resource A comprehensive guide and codebase for fine-tuning YOLOv9 on custom datasets, covering everything from data preparation to inference.
| Domain | Tools |
|---|---|
| AI & ML | Vertex AI, Gemini, OpenAI API, TensorFlow/Keras, PyTorch, YOLO |
| Cloud & DevOps | Google Cloud Platform (GCP), Cloud Run, Cloud Functions, BigQuery, Docker |
| Backend | Python, FastAPI, Node.js, Firebase |
| Frontend | React, Tailwind CSS, Astro |
- Hire Me: evanparra.ai
- LinkedIn: linkedin.com/in/eraphaelparra
- Email: admin@evanparra.ai