TECHNOLIFE

• Performed data processing and exploratory data analysis (EDA) for multiple business use cases.
• Designed & Implemented MVP recommender systems using collaborative filtering and content-based methods.
  • Developed back-end components with FastAPI and ExpressJS.
  • Exposed recommendation services via REST and GraphQL APIs.
  • Used MongoDB and Redis for storage and caching.
• Improved AI assistant features and optimized code efficiency.

SHATEL

• Designed and implemented LLM evaluation workflows (benchmarks, metrics, and purpose-specific tests).
  • Improved model evaluation reliability and comparisons, and reducing assessment time by 75%.
  • Implementing benchmarks from state-of-the-art research papers, enabling systematic evaluation of LLMs.
  • Developed custom benchmarks and adapted open-source ones to company-specific use cases, enabling more targeted evaluation.
  • Evaluated more than 60 state-of-the-art open-source LLMs and in-house models to guide model selection and improvement.
  • Compared and documented performance across diverse benchmarks to guide model selection and deployment decisions.
• Built data storage solutions and scalable pipelines to process large-scale datasets.
  • Reduced manual preparation time by 40%.
  • Improved reproducibility and documenting metadata.
• Deployed and monitored LLMs on Linux servers.
  • Investigated and analyzed different deployment settings, gaining insights on cost, revenue, and profitability.
  • Optimized resource utilization of GPU servers by 50%.
  • Ensured reliable system performance in production environments.

MATNO

• Developed and maintained scalable back-end services, serving thousands of requests daily.
• Designed and Implemented back-end services from scratch, enabling smooth integration with front-end and external services.
• Refactored existing legacy codebases, cutting technical debt and improving maintainability.
• Implemented and optimized REST APIs, using Node.js (AdonisJS, NestJS) and Python (FastAPI, Django) frameworks.
• Integrated relational (PostgreSQL, MySQL) & non-relational (MongoDB, Redis) databases into production-level systems.
• Deployed back-end services using Docker and CI/CD pipelines, ensuring reliable delivery and consistent environments.
• Developed automation tools using web scraping to streamline internal workflows and reduce repetitive manual processes.
• Implemented and deployed a Telegram bot to automate communication processes and improve service efficiency.

Data Science Research Laboratory

Supervisor: Dr. Mostafa H. Chehreghani

• Project: Improving Fairness in Recommender Systems using Regularization
• Researched fairness in recommender systems, focusing on GNN-based methods & regularization techniques.
• Developed PBiLoss, a novel loss function, improving fairness and personalization in multiple benchmark datasets.
• Preprint paper: PBiLoss: Popularity-Aware Regularization to Improve Fairness in Graph-Based Recommender Systems.

Game Theory and Mechanism Design (GTMD) Research Laboratory

Supervisor: Dr. Mohammad Hossein Manshaei

• Supervisor: Dr. Mohammad Hossein Manshaei
• Project: Converting Genome to Gene Expression in Cancer Cells with CycleGAN
• Investigated computational biology and GAN applications, using genetic data & CycleGAN.
• Applied DeepInsight to transform TCGA datasets into image format, enabling cross-domain translation.
• Conducted experiments on genome-to-expression mappings, contributing to understanding cancer cell gene expression.

• Search & Information Retrieval Graduate Course (Fall 2023, Fall 2024)
• Natural Language Processing Graduate Course (Spring 2024)

• Operating Systems Undergraduate Course (Fall 2021)
• Design & Analysis of Algorithms Undergraduate Course (Spring 2021)
• Basic Programming Undergraduate Course (Fall 2019)