Sheida Majidi, Software & Machine Learning Engineer

Portrait of Sheida Majidi, half illustrated and half photograph.

full-stack

Full stack engineer building fast, polished web apps, from UI to API.

<coder>

Software & ML engineer turning messy data into production ML systems.

  1. Intro

    Sheida Majidiis a softwareand machine learningengineer turningmessy data intoreliable products

  2. About

    About

    Based in Vancouver, BC

    • Python
    • SQL
    • C++
    • TypeScript
    • React
    • Next.js
    • GCP
    • BigQuery
    • Vertex AI
    • Snowflake
    • ML pipelines
    • ETL/ELT
    • Time-series

    Hi, I'm Sheida. I'm a software and machine learning engineer in Vancouver, currently building ML systems at FactualIQ. I like taking messy, real-world data and turning it into products people can actually rely on.

    Before this, I spent two years at IVADO Labs in Montreal building cloud data and ML pipelines for large-scale time-series and IoT data, plus a few LLM-powered features along the way. What I enjoy most is production ML: models that live inside solid software, not just in notebooks.

    I took the scenic route to get here. I started out in civil and geotechnical engineering (Sharif, IUST, then PhD research at Carleton) before switching lanes with a Master of Management in Analytics at McGill. That background still shapes how I work: own the problem end to end, and build things that hold up.

  3. Projects

    Selected work

    • Media intelligence2026

      AI news intelligence platform

      Part of the team building a product that turns a huge news stream into a daily, source-linked briefing for each organization.

      • Daily briefings drawn from 6,500+ publishers, public records, social and video.
      • Stories ranked, summarized and grouped into themes, each traced to its source.
      • Worked across the app: personalized feed, search, interests and admin tools.
      • Next.js
      • React
      • TypeScript
      • Tailwind CSS
      • MongoDB
      • Elasticsearch
      • Pinecone
      • AWS ECS
      • Terraform
    • Data platform & ETL2026

      Document scraping & enrichment pipeline

      The backend behind the news platform: our team's pipeline turning public web pages and PDFs into clean, searchable content.

      • Scraping with Playwright and httpx; PDF and HTML parsing with an LLM fallback.
      • LLM enrichment, embeddings and multi-stage dedup (hash → vector → LLM check).
      • Fan-out across serverless workers with Prefect: isolated sources, safe retries.
      • Indexed for search and retrieval, with end-to-end LLM tracing.
      • Python
      • Prefect
      • Playwright
      • LiteLLM
      • Gemini
      • Pinecone
      • Elasticsearch
      • MongoDB
      • AWS Lambda
      • OpenTelemetry
    • Digital health2025

      Wearable health insights & safer chatbot escalation

      On the team turning noisy wearable data into health insights people can understand and trust.

      • Daily production pipelines for FHIR wearable data, serving 150,000+ users.
      • An insights engine turning trends into plain-language health insights.
      • Chatbot escalation: sentiment plus frustration signals bring in a human.
      • Validation, monitoring and Terraform-managed infrastructure.
      • Python
      • Polars
      • BigQuery
      • GCS
      • Vertex AI Pipelines
      • Terraform
    • Retail2024–2025

      Energy optimization & predictive maintenance

      Part of a team building cloud pipelines that turn in-store IoT data into energy and maintenance insights.

      • About 2,000 stores, each with roughly 2,000 HVAC and refrigeration sensors.
      • Helped design sensor-to-equipment mapping with the Brick schema and RDF.
      • One ELT flow across four device APIs (two REST, two SOAP).
      • Airflow DAGs on BigQuery and GCS, from schema design to production.
      • Python
      • Pandas
      • Cloud Composer (Airflow)
      • BigQuery
      • GCS

  4. Where I am

    I'm based in Vancouver, BC, Canada.