About
I am a Senior Platform Engineer at Mojio Inc. I received my MSc in Computer Science from the University of British Columbia (2019) and my BSc in Computer Engineering from Sharif University of Technology (2016).
Education
University of British Columbia
Thesis: Machine learning of lineaments from magnetic, gravity and elevation maps
Supervisor: Prof. David Poole
Sharif University of Technology
Research in Machine Learning Lab
Supervisor: Dr. Mahdieh Baghshah
Publications & Talks
University of British Columbia
- Machine learning of lineaments from magnetic, gravity and elevation maps, MSc Thesis, 2019.
- A Convolutional Neural Network for semi-automated lineament detection, Computer & Geosciences, Elsevier, 2021.
- Deep Learning for Automated Lineament Mapping, TGDG Invited Talk, March 2021.
Sharif University of Technology
- Active Distance-Based Clustering using K-medoids, PAKDD 2016.
Industry / Independent
- Identifying blockchain-based cryptocurrency accounts using investment portfolios, arXiv, 2021.
- 3D Geophysical Inversion and Integration, EAGE Invited Talk, 2021.
- TorchMinusOne: A Tool for Deterministic Geophysical Inversion, Goldschmidt Conference, 2021.
Teaching Experience
University of British Columbia
- Computing Platforms for Data Science (DSCI 521)
- Probabilistic Graphical Models (CPSC 532R)
- Advanced Machine Learning (DSCI 575)
- Data Science Programming (DSCI 511)
- Internet Computing (CPSC 317)
Sharif University of Technology
- Digital Systems Design
- Signals and Systems
- Electrical Circuits
- Computer Structure and Languages
- Operating Systems
- Logic Design
- Fundamentals of Programming
Experience
Mojio
Mojio is a connected mobility platform processing large-scale telematics data from millions of vehicles across North America and Europe.
- Architected and developed scalable APIs on platform, hot-path message processing gateways & ingress, and distributed processing pipelines handling terabytes of vehicle telemetry data daily.
- Designed and implemented large-scale data analytics workflows to transform high-volume telemetry datasets into production-ready AI solutions for end-users.
- Leveraged and maintained different services including Event Hubs, Stream Processing, Cosmos DB, ASP.NET, and Kubernetes to build event-driven, containerized microservices and real-time processing systems.
- Led a large-scale data migration from Azure Cosmos DB to Couchbase, ensuring data consistency, performance optimization, and zero-downtime transition.
- Engineered and maintained video ingestion and live streaming systems, optimizing delivery pipelines for low-latency live streams, replay generation, and secure user access control.
- Built an automated reporting service powered by Retrieval-Augmented Generation (RAG) and AI models to generate periodic fleet efficiency reports, maintenance summaries, driver behavior analysis (road-score system)
GoldSpot Discoveries
GoldSpot Discoveries applies machine learning and data science to mineral exploration, integrating geophysical, geological, and geochemical data.
- Developed 3D models of subsurface survey data using stochastic modeling techniques and Gaussian Process optimization to improve accuracy and uncertainty quantification in geological exploration
- Designed feature extraction pipelines and implemented clustering algorithms on geological datasets using Deep Learning methods to enhance structural interpretation and data-driven decision making.
- Accelerated large-scale kernel computations and reduced memory footprint by implementing Lazy Tensors on GPUs using PyKeOps, enabling efficient processing of high-dimensional geospatial data.
Avestec Technologies
Avestec Technologies provides software solutions for geoscience data analysis, modeling, and inversion.
- Developed the full stack of the inspector platform, Avesoft™, for automated report generation and control of cameras and sensors on Avestec's SKYRON drone using Django.
- Enabled real-time transmission of sensor and image data across multiple protocols, including RS232, UART, XBee-900, and IEEE 802.11, ensuring reliable drone-to-ground communication.
- Designed, trained, and deployed a high-speed CNN model for live crack and defect detection on metal surfaces, running on NVIDIA's Jetson Nano embedded system mounted on the SKYRON drone.
Minerva Intelligence
Minerva Intelligence develops AI-driven tools for mineral targeting and exploration decision support.
- Assisted in developing machine learning models for mineral prospectivity mapping.
- Worked on data integration pipelines combining geological, geophysical, and geochemical datasets.
- Prototyped research ideas and supported senior engineers in deploying analytical tools.