cv

Education, research, and engineering experience. A PDF version is available from the button above.

Contact Information

Name Amir Ata Ghaffarian
Professional Title Research Intern, OIST · Machine Learning & Genomics
Email amirata.ghafarian@gmail.com
Phone +81 80-1984-3421
Location Onna, Okinawa, Japan

Professional Summary

Machine-learning researcher with an engineering background. At OIST I apply self-supervised learning and graph neural networks to genomic data; before that I worked in DevOps, SRE and backend engineering.

Experience

  • 2026 - Present

    Okinawa, Japan

    Research Intern
    Okinawa Institute of Science and Technology (OIST)
    Genomics & Regulatory Systems Unit, with Prof. Nicholas Luscombe and Dr. Charles Plessy.
    • Self-supervised learning and graph neural networks to decode genomic rearrangement patterns.
    • Multimodal representations of pairwise genome alignments to find structural variation at scale.
    • End-to-end HPC pipeline with Slurm and Singularity.
  • 2025 - 2026

    Tehran, Iran

    DevOps Engineer
    ICT Group
    • Deployed and troubleshot containerized services on Kubernetes.
    • Built monitoring and centralized logging with Prometheus, Grafana, Filebeat and the Elastic Stack.
  • 2025 - 2025

    Tehran, Iran

    Site Reliability Engineering Intern
    Neshan
    • Intensive SRE bootcamp: system administration, infrastructure automation, container orchestration and monitoring.
  • 2025 - 2025

    Tehran, Iran

    Research Assistant
    Center for Information Systems and Data Science, Sharif University of Technology
    With Dr. Babak Khalaj and Dr. Mohammad Hossein Rohban.
    • Predictive maintenance for batteries and bearings with Transformers and diffusion models.
  • 2024 - 2024

    Tehran, Iran

    Software Engineering Intern
    Yektanet
    • Dotanet ad server: ad retrieval, auctions and click/impression tracking in Go, PostgreSQL, Kafka and Docker.
  • 2022 - 2025

    Tehran, Iran

    Undergraduate Teaching Assistant
    Amirkabir University of Technology & Sharif University of Technology
    • Machine Learning (Prof. Ali Sharifi-Zarchi, Sharif University)
    • Advanced Programming, AI, Algorithms (Prof. Mehdi Ghatee, Amirkabir University)
    • Image Processing (Prof. Mostafa Shamsi, Amirkabir University)
    • Operating Systems, Software Engineering (Prof. Zahra Ghorbanali, Amirkabir University)

Education

  • 2020 - 2025

    Tehran, Iran

    B.Sc.
    Amirkabir University of Technology (Tehran Polytechnic)
    Computer Science
    • GPA: 4.0/4.0 (18.97/20) — last-year GPA 4.0/4.0 (19.60/20)
    • Relevant coursework: Machine Learning, Deep Learning, Artificial Intelligence, Data Structures & Algorithms, Information Retrieval

Honors and Awards

  • 2020
    Top 0.3% in the National University Entrance Exam (Mathematics)
    National Organization for Educational Testing, Iran

    Ranked 464th out of ~150,000 participants in the Mathematics track.

  • 2025
    Top 5% (ranked 3rd) among Computer Science students
    Amirkabir University of Technology

    Ranked 3rd among Computer Science students at Amirkabir University of Technology.

Interests

Research Interests: World Models & Reinforcement Learning, Robotics, Embodied AI, Large Language Models, Neuroscience, Computational Biology

Projects

  • Goal-Space Planning, reproduced (gsp-fourrooms)

    From-scratch tabular reproduction of Goal-Space Planning with Subgoal Models (Lo et al., JMLR 2024) on FourRooms, with exact subgoal models and an interactive report.

    • 8 experiments reproduced in about 3 minutes on one CPU core; GSP halves the episodes Sarsa(λ) needs to reach near-optimal (35 → 18).
    • Documents three open discrepancies with the paper, with candidate causes.
  • MyTorch — Deep Learning Framework

    From-scratch NumPy autodiff and neural-network framework with a PyTorch-like API: forward- and reverse-mode AD, optimizers, training diagnostics and run tracking.

    • Built with an agentic coding partner; every change validated by 1,200+ tests, gradient checks, strict type/lint gates and multi-version CI.
  • Knowledge Graph Generator

    Generates knowledge graphs from documents in Python, with FalkorDB for storage, Kafka for real-time processing, MinIO for objects and Redis for idempotency.

  • DiffRUL-CMAPSS

    Diffusion-based data augmentation (DDPM) with LSTM and Transformer models for aero-engine remaining-useful-life estimation, based on a 2024 RESS paper.

  • Search Engine

    Information-retrieval system in Python with TF-IDF, ParsBERT, K-Means clustering and FAISS.

  • Dotanet

    Scalable ad platform (Go, PostgreSQL, Kafka) with RESTful APIs for ad serving, auctions, and real-time tracking.

Skills

Programming Languages: Python, Golang, C/C++, Java
Frameworks & Libraries: NumPy, Pandas, TensorFlow, PyTorch, Scikit-learn, OpenCV, NLTK, FastAPI, Gin, Django
Databases: MySQL, PostgreSQL, Redis, FalkorDB, MinIO
Tools: Linux, Git, Docker, Docker Compose, Kubernetes, Slurm, Singularity, Kafka, Prometheus, Grafana, Nginx, GitLab CI/CD, ELK Stack
Languages: Persian (native), English (fluent)

Certificates