Arman M.
M.Sc. in ICT for Smart Societies
Experience
Oct 2021 - Dec 2025
4 years 3 monthsItaly
M.Sc. in ICT for Smart Societies
Polytechnic University of Turin
Relevant courses: Machine learning and neural networks, Programming for IoT, Big data for Internet applications
Aug 2020 - Feb 2021
7 monthsJunior Research Engineer – Part Time
Saipa Expansion Engineering Co. (SEECO)
- Led a feasibility study on adopting Automated Guided Vehicles (AGVs) in industrial logistics, mapping existing material-handling workflows and identifying opportunities to boost throughput and reduce manual intervention.
- Pioneered and simulated a machine learning based AGV route-optimization prototype, demonstrating a 20% reduction in idle travel distance and a 12% decrease in average task completion time.
- Deployed Python-based vendor benchmarking scripts, assessing 12 AGV manufacturers on 25 performance metrics, streamlining the selection process and reducing evaluation time by 40%.
- integrated statistical modeling and discrete-event simulation in Python to project ROI of 18% over two years by selecting top three AGV vendors for strategic trade partnerships.
Sep 2019 - May 2020
9 monthsJunior Structural Data Engineer – Part Time
Abadis Braces for Concrete Structures Company
- Tackled missed micro-cracks by training a Random Forest on sensors and load data, achieving 92% maintenance prediction accuracy and cutting unplanned downtime by 25%.
- Addressed material overuse through Python/MATLAB analysis of strain measurements, recommending brace design tweaks that reduced costs by 10% and extended lifespan by 15%.
- Eliminated manual reporting delays by building automated ETL pipelines and weekly dashboards, speeding decision cycles by 40% and boosting maintenance-scheduling accuracy to 90%.
Oct 2018 - Jun 2021
2 years 9 monthsB.Sc. in Electrical Engineering
Shahid Beheshti University
Thesis: “Machine learning approach for optimization of SLAM navigation via Extended Kalman Filter”
Relevant courses: AI and bio-calculations, Signal processing, Control systems, Embedded systems
Lorem ipsum dolor sit amet
DREAM-FLUTE: Commonsense QA Enhancement
- Integrated T5-BASE with the DREAM elaboration model, fine-tuning on 9,700 CoS-E samples, resulting in a 40.46% accuracy for multiple-choice QA and elevating to be one of the top solutions for the dataset.
- Augmented BERT with DREAM-generated elaborations to boost CoS-E accuracy to 20.48%, achieving an absolute gain of 0.09% and pinpointing areas for further optimization.
- Enhanced question-answering system performance by dedicating 10 weekly hours to code optimization, resulting in a 15% improvement in processing speed and a more efficient system overall.
- Technologies: Python, PyTorch, Hugging Face Transformers (T5, BERT), TensorFlow, CoS-E dataset.
Lorem ipsum dolor sit amet
ICT-Enabled Building Energy Optimization & Prediction
- Simulated a 12-zone school in EnergyPlus with hourly climate data, revealing heating, cooling, and electricity inefficiencies.
- Trained a surrogate ANN model on 40 design alternatives: applied NSGA-II optimization, reducing annual energy demand by ~19% through optimized glazing configurations.
- Automated EnergyPlus runs using Python/Eppy, processing >17,000 hourly data points to evaluate climate sensitivity.
- Emulated IoT sensors with MQTT, streaming real-time data to InfluxDB for smart building dashboards.
- Performed energy signature analysis via OLS regression, achieving R² up to 0.82, quantifying responsiveness to outdoor temperature.
- Configured LSTM neural networks predicting hourly heating/cooling loads with <5% error, enabling energy management.
- Technologies: EnergyPlus, DesignBuilder, Python, Eppy, scikit-learn, TensorFlow, MQTT, InfluxDB.
Lorem ipsum dolor sit amet
IoT Platform for Automated Irrigation & Plant Health Monitoring
- Constructed a data pipeline with Python, CherryPy, Flask, Paho-MQTT, Docker, and ThingSpeak to stream over 20,000 data points and leveraged those insights to optimize resource allocation and eliminate operational bottlenecks.
- Simulated soil-moisture, temperature, and humidity sensors with Python/Paho-MQTT, achieving 12 ms encryption and publishing latency over 1,000 messages (Fernet/AES).
- Engineered a Keras-based artificial neural network (ANN) using 201 data samples, achieving 94% accuracy in irrigation decisions, and optimized water usage by 15% compared to traditional methods.
- Developed a system that decrypted streams and detected leaf diseases via EC2-hosted Flask/InceptionV3 at 50 images/min, logging 100+ events, while personally managing 100 unique data logs daily.
- Technologies: Python, CherryPy, Flask, Paho-MQTT, Telepot, TensorFlow/Keras, scikit-learn, Docker, ThingSpeak, Fernet, SHA-256 hashing.
Summary
Final-year M.Sc. student in ICT for Smart Societies, specializing in Machine Learning, Computer Vision and NLP.
Design and deploy deep learning and RAG-based QA systems, anomaly detection pipelines, IoT solutions, and applying AI and ML techniques to engineering and construction challenges. Built end-to-end ML workflows—including AGV route optimization and large-scale time-series forecasting—with integrated MLOps practices.
Languages
Persian
NativeGerman
AdvancedEnglish
AdvancedItalian
IntermediateEducation
Oct 2021 - Dec 2025
Polytechnic University of Turin
MSc in ICT for Smart Societies · ICT for Smart Societies · Italy
Oct 2018 - Jun 2021
Shahid Beheshti University
BSc in Electrical Engineering · Electrical Engineering · Tehran, Iran, Islamic Republic of
Need a freelancer? Find your match in seconds.
Try FRATCH GPT More actions
Similar Freelancers
Discover other experts with similar qualifications and experience
- English
- Deutsch
2025 © FRATCH.IO GmbH. All rights reserved.