Projects | mmrahneshin

Predictive Maintenance of VRLA Batteries in BTS Systems

Fri, 01 Aug 2025 00:00:00 +0000

This undergraduate project was developed in collaboration with Nian Electronic Company to implement predictive maintenance for Valve-Regulated Lead-Acid (VRLA) batteries in BTS systems. The project implements the methodology from the research paper “Predictive Maintenance of VRLA Batteries in UPS towards Reliable Data Centers”using a Gradient Boosted Decision Trees (GDBT) model.

The system predicts VRLA battery degradation in advance for the company’s BTS systems, helping prevent unexpected failures and optimize maintenance scheduling. Key features include:

Battery Health Monitoring: Real-time assessment of battery condition using performance metrics
Deterioration Prediction: Early warning system for battery replacement needs
Cost Optimization: Reduced downtime and extended battery lifespan through intelligent maintenance

The implementation processes battery voltage, internal resistance, temperature data, and historical maintenance records to provide accurate predictions for critical infrastructure applications.

Pure Parity Seeking 2-3 Red Black Binary Tree

Sat, 15 Jun 2024 00:00:00 +0000

Implemented a 2-3Red-Black Tree based on the “Parity-Seeking” deletion algorithm from a recent academic paper [Link]. The project focused on reusability and object-oriented design in C++, structured as a library-like module.

This implementation provides a pedagogically sound yet efficient approach to deletion in 2-3 red-black trees, making it suitable for educational or competitive programming use.

Presentation: Competition-Level Code Generation with AlphaCode by DeepMind

Fri, 05 Jan 2024 00:00:00 +0000

Presented a technical overview of AlphaCode, DeepMind’s LLM for competitive programming, as part of the graduate-level Neural Networks and Deep Learning course. Covered its architecture, sampling strategies, and evaluation process.

Discussed AlphaCode’s use of large-scale code generation and filtering techniques to tackle programming problems, along with its positioning among human competitors in programming contests.

Age Prediction with Transfer Learning

Sun, 01 Oct 2023 00:00:00 +0000

Developed an age prediction model using transfer learning with VGG16 and PyTorch. The project was about fine-tuning layers, regularization through dropout, and preprocessing face images to enhance precision.

You can view the project on GitHub: Age Detection Using Transfer Learning.

Sepehr – Automatic Scoring System

Fri, 01 Sep 2023 00:00:00 +0000

Sepehr is an automatic scoring system developed for programming exercises, with a particular focus on data structures and algorithm design problems. It was initially created during my internship at Part Research Institute and later used in both university-level and national-level Dataleague programming contests.

The system automates the evaluation of code submissions, streamlining the process of grading and feedback for competitive programming and coursework. It has been actively used in:

Data Structures and Algorithm Design courses at Ferdowsi University of Mashhad,
The University-level Dataleague Contest (2023–2024),
And the National Dataleague Contest (2024–2025).

Sepehr is designed with modularity and scalability in mind, enabling flexible integration into different contest platforms and academic environments.

Seizure Signal Classification

Thu, 01 Dec 2022 00:00:00 +0000

Developed a seizure detection system using EEG data and machine learning models including SVM, Random Forest, and CNNs.

Performed preprocessing, feature engineering, and K-fold validation to ensure accuracy and generalizability. Focused on biomedical signal analysis in practical ML settings.

You can view the project on GitHub: Seizure Signal Classification.

Game Player Agent in Abalone Game

Wed, 01 Jun 2022 00:00:00 +0000

Developed an intelligent board game player, Abalone, utilizing Minimax with Alpha-Beta Pruning, which was coded in GDScript.

This was a project from the Artificial Intelligence fundamentals course and was about optimal search and decision-making in adversarial environments.

You can view the project on GitHub: Game Player Agent in Abalone Game.

AI Planning Framework

Fri, 01 Apr 2022 00:00:00 +0000

Created a planning system in JavaScript for classic AI problems like BlocksWorld and Monkey-Banana using heuristic and search-based methods.

Applied strategies like ignoring delete lists and forward/backward search to efficiently resolve planning tasks through state transition modeling.

You can view the project on GitHub: AI Planning Framework.