Shadi Akiki

About me

I am a Senior Data Scientist who builds turnkey data-oriented solutions. My highest value is in modeling of data relationships, but my work covers the full range of data acquisition, structuring, analysis, modeling, presentation, and alerting. My previous work has been for small to medium institutions in Boston, Singapore, Beirut, Dubai, Geneva, London, and Paris.

As of July 2019, my focus is on my startup, AutofitCloud Technologies, Inc.

Tools and Examples

Here are some of my active tools and example products that I’ve built

Field Tools Examples
Data science
  • LSTM autoencoder on economic data with feedforward network between latent variables
  • decision trees to split comma-separated lists with not all commas being splittable
  • reinforcement learning to design/operate a simulated water treatment plant
  • transcribe batches of videos with speech recognition with IBM watson
Data engineering
  • reconcile the same data from different sources to highlight mismatches for correction
  • scraping unstructured data from websites and reformatting into a structured format
  • postgresql (relational)
  • Mongodb (nosql)
  • Apache Spark (big data)
  • automated data pipelines between well structured sources or unstructured sources such as emails, pdfs, worksheets, etc
  • mirror data from a source software to a target unrelated software
  • version control of text exports of database tables to monitor changes in any field
Web frameworks/apps
  • django
  • flask
  • nginx
  • heroku (app hosting)
  • AWS EC2 (compute power)
  • platform for communicating well-formatted information between teams and track acknowledgements
  • platform for interacting with all above tools
  • platform for single point-of-entry of data entry that feeds to multiple different and unrelated software
  • git (version control)
  • linux (operating system)
  • bash (command-line)
  • (testing)
  • pytest (unit tests)
  • tableau (visualization)
  • implement new features in existing popular open-source tools and have them integrated into upstream codebase
  • Organize all code implemented in re-usable packages
  • Deploy platforms and automated tools in-house on existing hardware or in the cloud



Last edited on: 2019-11-29