Alireza Hashemi

Alireza Hashemi

Ph.D. Physics Student, Interested in interdisciplinary research on network & data science

Alireza Hashemi

Research Interests

  • Graph neural networks and machine learning on graphs
  • Complex systems regarding social and financial problems and network analysis
  • Applied machine learning
  • Data science, data analysis, and data reporting
  • Computational physics and simulation

Education

Work experience

  • blubank: Data Scientist (May 2020 - Aug 2022)
    • Fraud detection on user debit card transactional data using isolation forests, autoencoders, and feature engineering.
    • Training and stacking of several CNN architectures for face vs. ID card verification using transfer learning, on-device face liveness detection of live video.
    • Knowledge graph design and building, designed and developed a Neo4j database from several data sources to identify money-laundry networks, community detection for marketing purposes, fraud detection on money transfer, and recommender system for recommending merchants to users.
    • Marketing analysis & A/B testing of application features, high level presentations of growth and user usage data overview for managers and strategic planning.
    • Data engineering pipelines and automation DAGs using Apache Airflow and ETL with Python. experience with different databases: Oracle, PostgreSQL, MongoDB, MySQL, Clickhouse, Neo4j. Fact table design for OLAP purposes.
  • Rahnema College: Machine Learning Course Mentor (Mar 2021 - Jun 2021)
    • Mentored a small group of interns in machine learning for three months as a part of volunteer work and giving back to Rahnema College for my internship.
  • Rahnema College: Machine Learning Intern (Jan 2019 - Mar 2020)
    • 3-month training program in basics of machine learning. Final project: A recommender system for a music streaming service.

Skills

Python: NumPy, Pandas, Scikit-Learn, NetworkX, Keras, TensorFlow, PyTorch (and PyTorch geometric), graph-tool, C++: LAPACK, Computational Physics, Matlab, R, Basic Linux, SQL & NoSQL Databases, CUDA, Machine Learning: Supervised/Unsupervised/Semi-Supervised algorithms with standard libraries, Deep Learning: Image processing with CNN, Transfer learning, Graph neural networks, Autoencoders, Timeseries: Causal impact analysis, timeseries prediction algorithms for strategic planning, Anomaly Detection, Graph Databases: Neo4j, Teamwork Tools: Jira, Git, Confluence, Data Reporting: Metabase, Superset, Data Engineering: Apache Airflow, Python ETL, APIs

Publications

Talks

Other

  • NTD Hackathon runner-up team (covered in NPR.org here)

  • TA for Physics 1&2, Analytical Mechanics 1&2 (Sep 2017), Electromagnetism (Sep 2018), Nonlinear Dynamics & Chaos (Sep 2020)

  • Translation of the book “Dark Matter & Dark Energy” by Brian Clegg to Farsi

  • 23rd & 24th School on Physics, IASBS

  • Tehran on Complex Networks (TACN2018), Shahid Beheshti University

References

  • Prof. Amir Hossein Darooneh, Department of Applied Mathematics, University of Waterloo, Researcher at SickKids Hospital

  • Prof. Mohammad Reza Ejtehadi, Department of Physics, Sharif University of Technology

  • Prof. Moladad Nikbakht, Department of Physics, University of Zanjan

Hobbies

  • Movies, Hiking, Music