Jan Štihec Email & Phone Number

Stamford, Ct, Us

Zurich, Switzerland

We are the Open AI of closed expert knowledge domains - HSE, compliance, law, tax and insurance.With Artificial Intelligence (NLU, NLP, NLI, Generative AI based) products & services we are solving problems in Risk and Legal, Marketing and Sales, Operations, IT/engineering, HR for corporate clients.

Zug, Switzerland

Artificial Intelligence powered Enterprise SaaS for cost reduction and improvements in Occupational health and safety in Switzerland.

Client #1: Project: Analyzing Twitter communications during the 2021 OlympicsResult: Time series analysis of 417 accounts involved in the Olympics (over 250.000 tweets). The analysis involved sentiment analysis, deep emotions analysis, topic extraction, regression analysis of impact and impressions, flexible timeline analysis, interactive visualization.Tech: spaCy, statsmodels, numPy, pandas, D3.js, transformers, BERTClient #2: Project: Financial text summarizationResult: The solution provides an extractive and abstractive summary for the given input text. The model was finetuned on a financial dataset and is deployed as a REST API service.Tech: transformers, BERT, PEGASUS, spaCy, FastAPI, Docker.Project: PDF relevant information extraction, NER and entity resolution.Result: For each PDF the solution extracts predefined entities (name, address, email, number and date) and relevant sentences where relevancy is based on the underlying sentence topic/theme.Tech: Spacy, pyPDF2, transformers, FastAPI.Client #3: Project: Inventory item similarityResult: Developed an algorithm for calculating the degree of similarity between inventory items using Item ID and Item Description. For each item the algorithm returned 3 most similar items from the database (300.000 + items). The accuracy achieved was over 95% (measured on a smaller subset).Tech: scikit-learn, spaCy, pandas, numpy Client #4:Project: Collecting inventory items from a list of autoparts webstoresResult: Wrote a script for collecting inventory items from a list of webstores (15 webstores). Items were saved in a SQL database.Tech: Selenium, BeautifulSoup, MySQLProject: Geocoding a list of company addressesResult: Wrote a script which using different APIs converted addresses to geographical coordinates for several thousand companies.Tech: Selenium, pandas

Ljubljana, Slovenia

Project #2 (2020-2022) – Research team member*Research: Quantitative and qualitative analysis of unregulated parts of corporate financial reporting.*Description: My role included data collection, data cleaning, text pre-processing and tagging, vectorization, model building and evaluation, presentations. My main work was developing an algorithm for detection of forward looking sentences (FLS).*Results: We described part of our FLS detection approach in a paper available at: https://aclanthology.org/2021.fnp-1.4.pdf The paper was accepted for The 3rd Financial Narrative Processing Workshop (FNP 2021), which I attended and presented our work.Project #1 (2017-2019) – Research team member*Research: Analyzing the importance of informal forms of communication and unregulated part of annual reports on capital markets.*Description: My role included document parsing and sentence extraction, text pre-processing and tagging, vectorization, model building and evaluation, data interpretation and presentations. My main work was developing an algorithm for detection of semantic orientations in economic texts.*Results: We presented the algorithm in a paper available at: https://link.springer.com/chapter/10.1007/978-3-319-76941-7_64 The paper was accepted for The 2018 European Conference on Information Retrieval in Grenoble, France. Coworker and I attended the conference and presented our work.

Ljubljana, Slovenia

Helped a Ph.D. student with data preprocessing and developing an ML model for Sentiment Analysis.

Slovenia

Worked as a part of the team which executed an IT infrastructure upgrade (hardware replacements and software updates) at different Novartis locations in Slovenia.

Ljubljana, Slovenia

As a member of the board I organized social, cultural and educational events and worked as a student representative.

Bachelor'S Degree, Computer And Information Sciences, General

Master'S Degree, Business Informatics