Ali Abbas

At CreaVilla Horizon, I built and deployed CV-Screening, an AI recruitment platform with a Python/Flask API on Google Cloud Run. It extracts candidate data from PDF and Word CVs and scores them with the Gemini API.

I made 200-CV batches recover from restarts using checkpointed Cloud Tasks slices and a Cloud Scheduler sweeper. I also added a rule-based scoring engine as a second opinion and fallback, plus a 60-CV regression corpus to test name extraction.

For security and cost control, I added emailed sign-up codes, rate limits, TOTP two-factor authentication for admin accounts, and a budget control that pauses AI at 100%. I moved Cloud Run to scale-to-zero with request-based billing and fixed production issues involving stale revisions, IAM/CORS gaps, and empty LLM responses.

On QuantFlow, my AI quantitative trading platform project, I worked with FastAPI, PostgreSQL, Celery, and Next.js across market data, forecasting, and risk controls. The project has 960+ tests, and I found and fixed a bug that caused marketable limit orders to execute at their limit rather than the market price.