Hassan Nasr Hassan

I build production AI and data systems. Currently shipping real-time Arabic voice AI at Arcain. Previously 4+ years of data engineering across Canadian federal agencies.

Featured Work

Experience

  1. AI / Voice Engineer · Arcain

    Dec 2025 — Present
    • Fine-tuned Whisper Large-v3 with LoRA on 53k Arabic samples across 14 dialects. Cut WER 30.6 → 17.8% (41.8% relative) and p50 ASR latency to 345 ms.
    • Shipped a dual-mode FastAPI voice service with streaming ASR / TTS over WebSockets, Silero VAD v5 barge-in, and OpenTelemetry on AWS GPU.
    • Built an Emirati Arabic TTS fine-tune on the F5-TTS architecture with a zero-downtime env-var rollback pattern.

  2. Data Engineer · Innovation, Science and Economic Development Canada

    Aug 2024 — Present
    • Built incremental and full-load ETL pipelines (Python, SQL, AWS Lambda + EventBridge, EC2) moving data across RDS, PostgreSQL, and Oracle.
    • Implemented validation logic and stored procedures supporting 10+ analysts on 4M+ record datasets. Cut validation effort 50%.

  3. Data Engineer · Transport Canada

    Feb 2023 — May 2024
    • Built automated Oracle-to-Databricks pipelines with PySpark, curating data across 945 projects for 20+ program managers.

  4. Data Analyst · Fisheries and Oceans Canada

    May 2022 — Feb 2023
    • Automated cleaning and matching of ~2,000 raw survey responses and built visualizations surfacing trends across Atlantic and Pacific fisheries quota data.

  5. Data Scientist Intern · Defence Research and Development Canada

    Jun 2021 — Dec 2021
    • Built an OCR pipeline recognizing typed, handwritten, and printed Arabic text, adapting prior OCR research to a multi-format task.

  6. Software Engineer Intern · Statistics Canada

    Jan 2020 — Aug 2020
    • Built a user management module for the Census 2021 application in C# / ASP.NET / Azure handling names, roles, and groups.

Education

B.Eng., Software Engineering · Carleton University

Sept 2017 — Jun 2022