* Designed and developed a microservice architecture for a General AIOps system POC.* Developed Data Ingestion and Data Analytics services using Python, Kafka, InfluxDB, and MongoDB* Developed tools to detect anomalies in largescale Time Series* Developed a library to query SQL databases with a PromQL like language usingPython and LARK Parser (https://github.com/adavoudi/PromSQL) * Developed tools for the Extraction and classification of entities relations from natural text using a rule-based and deep-learning based methods* Performed graph analytics such as Link Prediction, Node Similarity (using embeddings), and Shortest Path using Neo4j

* Developed a framework to model Bayesian Networks for anomaly detection (https://github.com/adavoudi/bayesnet)* Developed several algorithms to detect multi-dimensional time series anomalies* Developed several data ingestion pipelines to collect, transform and store all the incomming data in the database (using Kafka and InfluxDB)* Developed the Backend GraphQL API using Django and Graphene

* Trained an image tagging deep model based on the company's data* Optimized the model to run on CPU using Knowledge Distillation

* Trained several deep learning models for different purposes such as Fire and Smoke detection and localization, car localization and classification, weather detection, ... using Tensorflow and PyTorch* Designed and developed a framework to train an end-to-end deep license plate recognition model by only giving it the template and font of the target license plate (the system could generate training samples based on the given template).* Developed several computer vision and image processing APIs for the company’sproducts using C++, and OpenCV* Implemented a framework for surveillance camera monitoring. The pipeline includeda video management system (using python, FFMpeg, MongoDB), an Object Detection and Tracking System (using Tensorflow, Tensorflow Serving, Kafka, Python), a complex event processing system (using apache Flink and python), and finally an alerting system.