Пожалуйста, войдите, чтобы скачать резюме.
Experience level:
3-5 years
Education level:
Languages:
Russian: Native
English: C1
Serbian: A1
Conducted mobile network kpi analysis using base station adjacency data
Implemented error prediction models on mobile network using probabilistic models
Designed and implemented Airflow-orchestrated tasks for anomaly detection in mobile network traffic using
unsupervised learning
Designed and implemented Airflow-orchestrated tasks for kpi time series prediction, achieving reasonable accuracy
on hourly data, while preserving computational resources by using effective models
Conducted error analysis and prediction with underlying reason consideration
Using transport network device adjacency, pinpointed exact hop causing packet loss problems
Migrated company’s IP address management system from an Oracle database to NetBox
Parsed router configs, organized data into Netbox
Analyzed IP Backbone crash data, including trouble tickets and events from the crash database, created a dataset
of most frequent problems
Conducted connectivity analysis using graph neural networks to identify weak points in the core network
infrastructure
Analysed and predicted spreading of crash influence using graph neural networks. Calculated probabilities for
possible device crashes up to a week in advance
Predicted location of optic fiber cuts with classic ML algorithms using DWDM crash info
Created a dynamically-updating view for bot statistics in google sheets via SheetsAPI
Done research on Bayesian optimization hyperparameter tuning. Improved backtesting time and results by 33%
Worked in a team on a python library for minute trading bot functionality
Done overall analysis of latest financial data prediction models
Languages and instruments: Python, SQL, Gitlab, Docker, Postman, Wireshark, Zabbix, Airflow, Grafana
Libraries: NumPy, Pandas, requests, PyTorch, TensorFlow, Scikit-learn, OpenCV, Matplotlib, pySpark