About me

A computer engineering graduate student who is interested in data science and computational mathematics.Having worked on new ML techniques like FL in IoT and Advertising tech. Dedicated to software engineering, having more than 2 years of expertise, and specialising in back‑end web development, software architecture and patterns, and machine learning.

I was actively building up an advertising technology startup, dealing with different retargeting technologies and I came across the novel approaches of privacy preserved advertising, like the usage of Federated Learning. Also, I have worked on the ways of distributed software scaling and I'm completely familiar with Kubernetes HPA, Reactive Programming Patterns, Distributed Database Management and Event Driven Programming via the usage of Spark Streaming.

What i'm doing

  • design icon

    Linux System Administration

    I had kernel compiled my own distro when I was at high school :D That project got nominated in 2015 World Summit Awards for a world competetion

  • Web development icon

    ML/AI Predictive Analysis

    I have used RF regression algorithms for predicting critical temperature, while optimizing the model hyper parameters and doing K-folds for cross validation. The hyper parameters where based on two search approaches, grid search and random search. The result will be published in a paper within the next month. Also, I have done Wavelet/Signal processings and I'm interested in Attention-LSTM

  • mobile app icon

    IoT

    As the new advertising constraints by google and apple has been announced, I was interested on researching the ways of improving advertising targeting accuracy by the usage of federated learning. FL and its usecases, especially asynchronous and semiasynchronous approaches within stochastic gradient descent grabbed my attention very much. Hence, a paper concerning the usage pf FL in predictive modeling will be published within the next month.

  • camera icon

    Ad Click Probability Prediction and Markov Chain for attribution modeling

    I have done a predictive modeling for analyzing the probability of click per each user. The whole process consisted of a decision tree, for checking the most related ads to each user,then publishing the bucket into the Elasticsearch. Whenever the user sent a Bid Request (ORTB 2.5), the ad will be served and based on the CTR history and the user behavior,I have researched the usage of Markov Chains for attribution modelings.

Resume

Education

  1. Azad University Science and Research Branch Tehran

    2021 — 2023

    Master of Science -- Computer Engineering, GPA 3.8/4.0, Courses: Parallel Computing, Advanced Algorithms, Seminar, Advanced Software Engineering, Advanced Networks, Distributed Computing, Advanced Software Patterns, Software Performance Analysis

  2. Azad University Central Tehran Branch

    2016 — 2021

    BSc of Computer Engineering GPA : 3.45/4.0 Selected for Industrial Cooperation with the most prestigious startup accelerator in Iran among 400 students omnis..

Experience

  1. Amirkabir University Innovation Center - StartUp Camper

    2020

    As our team got accepted out of many participant teams, we launched a referral marketing automation platform with the goal of managing affiliate links and automating commissions and gifting processes (Like Referral Candy, We Gift, CJ, etc.) ‑ Built and lead a team of 4 people to develop, and support a real‑time marketing automation platform ‑ Managing the deployment and architecture design for supporting high concurrency ‑ Created a complete CI/CD pipeline for the development team ‑ Managed a fault tolerance deployment with K8s HPA ‑ Enhanced RDBMS performance for handling A/B/n referral links

Contact

Contact Form