Huawei Russian Research Institute:Researching optimization algorithms related to Large Language Models (LLM).

Developing machine learning models for passport recognition system (R&D project).Main methods: neural networks using Python.Used libraries: TensorFlow+Keras, OpenCV, NumPy, Matplotlib, YoloV5.DB: Oracle, PostgreSQLI proposed the main pipeline for the passport recognition system and trained models for severalkey modules: object detection of input image, orientation detection.I adopted Dlib machine learning toolkit written in C++ to Java using JNI to solve fraud recognitionproblem. The adopted library supports multithreading. The main purpose of this library is torecognize and compare faces.My main tasks:* Train neural networks to solve given problems.* Assess trained neural network models.* Find ways to improve recognition metrics.* Test hypothesis.* Analyze obtained results.On other company projects:* Made several models which predict the probability of subsequent purchases for a customer(scikit-learn).* Developed microservices using Python (asyncio).Since 2020 I have been learning machine learning in the Artificial Intelligence University.I strengthened the following knowledge and skills:* Convolutional neural networks, recurrent neural networks.* Object detection.* Image classification.* Image segmentation.I used these skills in the current project. This allowed to implement all pipeline modules and runthe passport recognition system in production.As a Senior System Analyst:* Designed architectural solutions for systems integration.* Implemented various integrations with banks: REST API, SOAP, IBM WebSphere MQ;internal integration - Rabbit MQ.* Designed data flows, integration formats and solutions.* Participated in functional and load testing.* Wrote complex SQL queries (Oracle).* Develop scripts for mass operations (Python, Pandas).

Key responsibilities:* Leading team of analysts.* Writing project documentation.* Developing bank information systems.* Collecting, analysis, formalizing and agreeing customer requirements for the projectedsystem.* Leading tasks from analysis to deploying to production.* Building communication between team members.* Communication with customer.* Participating in project planning.

Key responsibilities:* Participating in developing architectural solutions.* Developing software which provides data integration between systems and forms reports.Stack: C#, .NET Core, PostgresDB, MongoDB, AngularJS, Docker, RabbitMQ.To create high load system modules, asynchronous programming has been used.A unique access control mechanism was implemented in the system. It enabled flexibleopportunities to control user access to the system and protection from unauthorized access.

Key responsibilities:* Conduct initial assessment of IT maturity of distributor. Identify gaps and contribute development of deployment plan together with business analyst.* Train IT staff with interfaces specification. Provide ongoing control and support of interface development and manage integration testing.* Manage bug fixing by means of working with distributor IT personal and iSales vendor if required.* Make sure that distributor orders IT equipment in advance to secure go live date and ensure that ordered equipment fit PI equipment list. In case of DS3 distributor provide demand to IT Rollout Lead.* Make sure that distributor’s network meet PI specification otherwise raise issue and provide recommendation to distributor to update network. Manage network update. If network update is not possible due to any cause make sure that both businesses formally accept risk of nonconformance of network to PI requirements.* Do initial system configuration (mobile setup, security configuration, etc). Provide training of IT staff of distributor (integration approach, equipment configuration, security) and contribute business deployment analyst with training of business team (back office and mobile application trainings).* Manage risks and issues by means of timely identification and proper communication. Ensure timely and proper escalation if required.* Agree and arrange distributor go live together with business analyst and distributor integration manager (open items (if any), resources, go live cutover plan, etc).* Execute handover to support according to sustain acceptance criteria.Achievements:* Successfully deployed PepsiCo CRM to 12 distributors. Each deployment included business-processes and IT systems architecture analysis, developing a deployment plan, controlling deployment quality and deadlines.* Developed detailed system description for CRM developers on 60 pages in English.* Trained 100+ sales representatives to work with PepsiCo CRM.

Key responsibilities:* Develop information systems.* Write system specifications for technical specialists.Achievements:* Came forward with an initiative to adopt the system that provides the ability to load texts with grammatical markup into the Internet in automatic way. This reduced the charges of Institute by 300,000 rubles in a five year perspective.