● Responsibilities: o Contribute on improvements of Search systems of the biggest online retail company in Romania – cutting edge research on statistically efficient methods for estimating Position Bias – using state-of-the art methods, estimating Search Relevance scorings, Learning to Rank, Search Query Auto-Completion. Performing Data Science research of the state-of-the art papers in the field of search ranking and Query Auto-Completion and implementation of the new methods with further evaluation with A/B tests in production. Contributing with Software Engineering expertise on design & implementation of the new flows of massive data coming from web-logs.

● Responsibilities (brief description as NDA is very strict): o Work on Software Engineering & Data Science related parts of projects; Contribute on development of a couple of projects for internal EC customers o designing & developing AI-related PoCs (mainly NLP methods), effort estimations, consult&help connected teams, design of the new systems running in distributed environments and actual hands-on development of internal projects

● Responsibilities: o Main responsibility is supporting pre-sales initiatives, designing & developing AI-related PoCs for existing and potential customers, effort estimations, as well as consult&help customers solve their business pain points using modern AI technologies for time series forecasting, Natural Language Understanding, Computer Vision, DeepLearning and shallow ML. o Perform Data Science and Data Engineering trainings for Iasi employees, offer support and supervise Data & Analytics components of projects located in Iasi.

● Company activity field: Apollo Technologies is a Romanian AI start-up developing their main product - Apollo Chatbot service, running on Facebook platform. The service it develops is intended to be set up by restaurants, with customized menu, and posted on the restaurant's Facebook page for receiving online food orders and table reservations without human intervention. Apollo Technologies got funding from Romanian Government. The start-up failed due to the lack of good sales and impossibility to acquire enough clients.

● Company activity field: Factmata offers an artificial intelligence platform which strives at reducing online misinformation. It aims to score many signals like Hate speech and abusive content, Propaganda and hyperpartisan content, Spoof websites and fake news, Deceptive or misleading content, Extreme clickbait content, Aggressive rhetoric and style, Biased stance and arguments and Invalid claims. Factmata have developed patent-pending technology to provide a real time quality, safety and credibility score to any piece of content on the web. The Factmata start-up got fundings from Google, Mark Cuban, Biz Stone(Twitter cofounder), Craig Newmark(Craiglist), Mark Pincus, Ross Mason, etc, ● Contributions: Collaborated with Factmata team and associated professors (Sebastian Riedel, UCL and Andreas Vlachos, Cambridge) on implementing new Natural Language Processing & Machine Learning techniques for detecting Fake News, Political bias, Hate-speech, Racism, Sexism and Terrorism. Integrating ML & NLP algorithms within Factmata detection systems.

● Responsibilities: o Crafting the right architecture to solve the problem, make high-level design choices, assure the system scales to the future business needs, document and communicate solutions, make sure everyone is using it, and using it correctly. Resolve technical problems, resolve disputes and make tradeoffs, plan for new technology insertion, manage risk identification and risk mitigation strategies associated with the architecture. Designing and implementing PoCs for new clients and contributing to RFPs implying Artificial Intelligence technologies. ● Some of contributions: o design and oversee implementation of secure integrations and data exchange with 3rd party services like Zapier.com for Pure360. o design horizontally scalable distributed storage and data flow architecture that fits new business requirements for the Pure360 (www.pure360.com) email service. o elaborate performance testing strategies, design architecture of new performance sensible components of backend systems for BetterWorks (betterworks.com). o performing AI technologies related RFPs and implementing PoCs in areas of NLP & ML ● Technologies used: Python, Java, Scala, ML (scikit-learn), Deep Learning (Tensorflow/keras), NLP(NLTK, Stanford CoreNLP, SpaCy, Wordnet), AWS (S3, EC2, SageMaker, AML, EMR, Textract, Kendra, Comprehend, Forecast, Rekognition, SQS, RDS, VPC, Gateway, SQS, Dynamo, Amazon ElasticSearch Service, RedShift, CloudWatch), SnowFlake on AWS, GCP (BigQuery, DataFlow, BigTable, DataStore, PubSub, Natural Language, DialogFlow) , Microsoft Azure (HDInsight Spark, DataFactory, VM, Function App, DataFactory, Managed SQL, Synapse, DevOps, Blob), Web Scraping/Web Crawling, ElasticSearch, Containerization related technologies like Docker, Kubernetes

● Responsibilities: o Being a mix between Data Scientist, Machine Learning Engineer and Software Developer, all of these applied on Big Data and highly distributed systems. o My main target is to find and implement engineering solutions on defined business problems affecting availability, data protection and security of Amazon public websites, as well as research, develop and integrate new methods to improve automated learning ability and quality of the Amazon Automatic Robot Detection services. Make them faster, smarter and more adaptable to changes in robotic and human behavior. I use data science and statistical learning algorithms to make machines learn fast and become better than we humans could do on discriminating between robotic and customer web service requests.

● Responsibilities: o Design, maintain and fix architecture deficiencies of services owned by Robot Detection team running on small and medium-size clusters(1K-10K machines). o researching and developing new detection methods and models, train supervised machine learning models that successfully detect robots, writing distributed processing pipelines in Scala, Python for Apache Spark, Java for Apache Hadoop Map Reduce and PIG Latin for Apache PIG o Design and develop layouts and process schemes for large-scale and complex data sets, manage the release process, establish application deployment process and supervise deployment to staging and production servers o maintaining contact with our clients through trouble ticketing system and making necessary investigations of robots o Solving complex problems, applying appropriate technologies and best practices. Splitting the work into parallel tasks that can be performed and then reassembled successfully. Work with my team to invent, design and build software that is stable and performant. Training new team-mates on how company’s software is constructed, how it operates, how secure it is, and how it fits into the bigger picture.

● Company activity field: AIDevel is a small privately-held Romania based company committed to offer high quality work in several areas, like NLP, Machine Learning, Web Robots. ● Responsibilities: My responsibilities were to assure the successful execution of the company’s business mission through development and deployment of the company’s projects within two company offices from Romania - Bucharest and Iasi, collaborate with the potential customers to develop use cases and specifications that describe the implementation of the company’s services, select and manage company staff, establish and supervise the software development process, setting objectives and assessing progress as defined by the selected software development methodology, review and approve proposed development releases, maintain upto-date knowledge of technology standards, industry trends, emerging technologies, and software development best practices by attending relevant conferences and reading widely, share knowledge, mentor, and educate the organization’s management and staff with regard to the company’s technological vision, opportunities, and challenges, supervise recruitment, training, retention, and organization of all development staff in accordance with the company hiring process, personnel policies, and budget requirements, as well as actively participate at projects technology research.

● Responsibilities: leading a small team of 5 programmers, interviewing, hiring, mentoring and training new team members, as well as achieving project development goals established by the company management. Contributing to technical design and development of the real-time data collection systems, collecting Level2 order-book data in from multiple data providers in parallel (IQFeed, PFG, Interactive Brokers), storing the level2 data in an efficient way fr retrieving for backtesting system, desing an efficient backtesting system for strategies hypertuning and elaborating new trading strategies using Machine Learning algorithms; Supervising real-money acount trading robots running in real-time. ● Achievements: Delivered a low-latency, real-time high frequency automatic trading platform, trading simulator for “backtesting”/evaluating profitability of trading algorithms that was successfully launched and used for algorithm profitability evaluation, designed and developed trading algorithms, pattern detection techniques and training machine learning models. Interviewed, hired and successfully integrated 3 new team members.

● My responsibilities were researching and applying Natural Language Processing techniques to extract various information from text and use that data to train grading models using supervised machine learning algorithms to achieve better results in paper grading module.

● Working on various projects with clients from all over the world ● Some of the projects: several small projects with developing of metaheuristic and evolutionary data-mining algorithms, a dozen of web-crawlers, web-scrapers and other web automation applications, developing parts/modules of TCP/IP based communication protocols to operate with NASDAQ and CME stock market brokers.

● Main responsibilities and activities: malware analysis, development of decryption & detection methods, prevention and cleaning of active malware, elimination of malware from networks and systems of company clients. ● Working in shifts: offering technical support, solving malware related cases on costumers’ systems, maintaining internal/external systems online, 24/7 services ● Automatic collection of malware from various sources ( in order to have fresh samples in the collection ) ● Creating and testing of detection and disinfection solutions against different types of malware using heuristic and machine learning techniques ● Development of internal tools with C++/Perl/Python, deployment and maintenance of multiple database management systems, design, development and maintenance of distributed systems for Anti-Malware Lab and automated malware decryption & analysis purposes. ● Internal testing different products for major tests in the industry ( AV-Test, AV-Comparatives, Virus Bulletin ) ● Writing stand-alone or product built-in disinfection routines for active threats ● Developing data mining scripts and adaptive web crawlers in order to build a whitelisting system and other generic routines for handling false alarms of the product. ● Sample exchange with different individuals or organizations ● Technologies used: Cryptography, Reverse Engineering, Malware Analysis, Disassembling, Advanced Networking, Hardware Emulation, Distributed Systems, Web Scraping/Crawling, Intel x86 Assembly Languages,C, C++, Perl, Python, Delphi, Java, Microsoft SQL Server, PostgreSQL, MySQL, Django, Machine Learning (Supervised & Unsupervised), Advanced heuristic/metaheuristic algorithms(Genetic Programming, Markov Models, etc..)