Alex Paddubitski

Project: Segmenter

Solution for segmenting Automatic Speech Recognition (ASR) results into media subtitles and audio scripts

.NET, Azure (WSS and HTTP API, OAuth, Azure Storage, Web Apps, App Insights, Dev Ops, etc). Remote team of 3, France

• the only backend developer in the team. Developed almost the whole product backend
• backend API development, design, maintenance and monitoring, etc. literally all the backend area is on me
• developing custom algorithms
• collaboration with ML and NLP experts, integrating custom NLP models into product

Project: StudioNEXT (Virtual Studio)

Solution for collaborative video dubbing outside a professional studio with minimal tech skills and equipment

.NET, Azure (WSS and HTTP API, OAuth, SQL, Azure Storage, Web Apps, App Insights, Key Vault, Dev Ops, etc). Cross-platform client app (Angular, electron) Remote team of 3, France

• the only backend developer in the team. Developed almost the whole product backend
• backend development, design, maintenance and monitoring, etc. literally all the backend area is on me
• the team highlighted by the company’s monthly merit recognition initiative for quick reaction to customer needs and development ahead of the department’s infrastructure
• the team is participating in the next-gen standard development and early adoption collaborating with a global top media company

Project: StudioNEXT (Virtual Studio)

Started as a contractor, continued as fulltime employee in Transperfect

Mostly outstaffing for a large localization provider

Project: Machine Translation Quality Tracker (MTQT)

Solution for machine translation services quality evaluation. Helps to make best MT choice for particular domain, customer, type of content. Constant reporting of quality trends, used for feedback to Google, Bing, Amazon, DeepL and other MT providers

Azure Functions, Azure Storage, SMTP, multiple MT API connectors, adopting and implementing text comparison and analysis algorithms. Architecture adopted for background daily\weekly execution, sharing production MT accounts and distrubuting each test run in time not to affect much throttling limits. Remote team (management), Spain

• Solo developer, implemented based on a demo piece of code from an internal conference
• CD\CI, development, maintenance and troubleshooting

RnD team

Broad range of tasks using .NET\C#, python, Azure. Utilizing various NLP technics starting from normalization (stemming, lemmatization), including but not limited to min-hash indexes, word embedding, tf-idf, n-gram based analysis for tasks like language\domain detection. Remote team of 4, Spain

• Developing automated QA methods for localization
• Batch processing company’s data for business questions (similar do data science tasks), using Azure Storage and Azure Batch capabilities
• Building text indexes for various analytical purposes
• Other tasks related to mining, warehousing and batch processing textual data

Project: Linguistic Toolbox (LTB)

Solution for automated text localization QA, integrated with user-facing systems Transformation of an old desktop tool to client-server architecture based on Windows services for running batch jobs, “greedy consumer” pattern + AMQP (Rabbit), ASP.NET Web API and Basic monitoring Web Portal (ASP.NET MVC). Remote team of 2-3, Spain

• Started as casual side resource to help with transformation and troubleshooting
• Finalizing the transformation stage and becoming solo backend developer of the system
• Refactoring and stabilization after app transformation phase
• General feature development (both server infrastructure and functional features)

Project: Translation Management System (TMS)

Solution for workflow orchestration to join manual localization process tasks with automations provided by ~30-40 internal and external systems. Heavily based on connectors and .NET Workflow Foundation. 10+ years in prod, 1TB SQL DB, x2 load each year.

~5 years. Distributed team (USA, France, Belarus, India) of size varying from 10 to ~30 people. Started as a middle software engineer, quited as the Belarusian subteam teamlead role. Scrum-like dev process

• General development and fixing bugs
• Planning, time and complexity evaluation for the team’s scope
• Working on system design solutions
• Refactoring legacy code and help with refactoring plans for the team
• Constant SDLC practices refinement (QA automation assistance, GIT usage model, QA environments maintenance, CD\CI automation, quality control automation, etc.)

Other activities (on-site, for Itransition)

• Technical interviews for .NET Software Engineer positions
• Mentorship for junior team members
• Knowledge sharing collaboration: organized and ran informal meetups for improving Algorithmic and Data Structure Skills
• RFX and RFI processing for the company
• Technical coordination experience for an internal project (the company’s employee management system)