Experience

Mithra-Ai Solutions

• Netherlands
• Software Development
• 1 - 100 Employee

• Machine Learning Engineer

  • Oct 2022 - Present

  - Developed a procurement classification model capable of handling both zero-shot and supervised learning scenarios, optimizing accuracy for reviews by users while also accommodating cases without reviews. - Enriched supplier information by creating a streamlined pipeline using the Google Search API in conjunction with Chat GPT, resulting in enriched and up-to-date data for improved decision-making. - Demonstrated expertise in architecting Vertex AI and Kubeflow pipelines for executing diverse classification, clustering, and data enrichment tasks, driving efficiency and insights across the organization. - Designed taxonomy suggestor algorithm given procurement data sets for various clients, providing valuable recommendations and insights to optimize procurement strategies. Show less



Charisma Financial Group | گروه مالی کاریزما

• Iran
• Financial Services
• 200 - 300 Employee

• Machine Learning Engineer

  • Sep 2022 - Present



Tadbirgaran Farda

• Iran
• Financial Services
• 1 - 100 Employee

• Machine Learning Engineer

  • Jan 2018 - Aug 2022

  - Implemented ETL process using Airflow to collect and process raw market data from multiple sources, reducing execution time from 5 to under a minute. - Developed processes for merging, cleaning, and validating data for over 1400 tickers on the Tehran Stock Exchange. - Designed database architecture for a wide range of market data using SQLAlchemy ORM. - Performed exploratory data analysis on market data to understand the data structure, identify outliers and anomalies and quantify relationships between variables. - Developed a module for feature extraction from 113 technical indicators and 20 fundamental indicators by auto-encoders using Tensorflow. - Developed a deep reinforcement learning portfolio manager for an ETF, which is in the top 5 of the best ETFs in Iran in terms of return. - Implemented custom dashboards to view portfolio management system’s predictions and performance using Flask and Plotly Dash. - Created a custom cross validation module in order to train and test different machine learning models in the time series data and perform precise backtest. - Built a customer segmentation model using k-means clustering with turnover and asset data, which helped marketing team gain insight about customers for designing different products. - Configure CI/CD pipelines so that the project gets tested with every commit and deployed to production server. - Dockerize previous and new projects in order to make deployment process easier. Show less