Introducing the No Hiring Fee Package!
Post vacancies
Hire without commissions
COMPUTER SCIENCE - 23201 Outstaffing
Experienced in solving applied artificial intelligence problems keeping in mind the integration of algorithms as software components to solutions workflow. Worked on the integration of developed methods and algorithms using REST services. Practical knowledge in handling uncertainty in data, solving heuristic search problems, and algorithm parallelization using Map Reduce technologies. Experienced in text and opinion mining workflows applied to text categorization, opinion polarity calculation, and unsupervised topic detection. Specialized in supervised classification algorithms, feature selection problems, and knowledge representation.
■ Projects:
□ CubanEngineer platform: Is a talent pool and recruiting platform for software engineers that resumes the user information spread in several sources on the internet like GitHub, GitLab, Stack Exchange, among others. The main job with the team was to develop a skills processing module that allows evaluating skills, infer new skills and get insights from provided data. This feature increases the product value for engineers who want to become better and recruiters who want the best engineers on the job market.
Duties:
■ Develop an engineer's skills evaluator using GitHub, GitLab, and Stack Exchange data
■ Propose a pipeline that allows inferring new skills from data provided by the APIs using a fuzzy cognitive map as the main algorithm for the evaluation process
■ Create a well-documented API for the evaluation service using Flask
■ Propose a graph-oriented data model to store the knowledge using ArangoDB
■ Define a neural graph structure for the knowledge associated with the skills
■ Projects:
□ DimeTu: Is a module to process opinion-related data used by the customer service department. The module was designed to find the main issues and flaws in company services and also evaluate how customer's opinions are related to provided services, improving in a short time the general opinion state and the quality of provided services.
Duties:
■ Select a state of the art method for polarity estimation and an unsupervised topic detection method. The selected method is described in this paper
■ Implement a semantic approach to find the main topics from a set of opinions using a hierarchical agglomerative clustering
■ Implement a method to estimate the polarity from a set of opinions using a fuzzy aggregation operator and spaCy word embeddings
□ Damix: Is a business intelligence platform that offers several advanced analytics. It includes reporting, data visualization, and BI predictive analytics. As a data scientist, his main job was focused on business intelligence and predictive analytics, proposing and developing the methods to provide evaluation, prediction, and simulation insights using large amounts of mixed and incomplete data. These features increased the number of clients by 12% and sales by 21.3%
■ Projects:
□ SOCACI: Is an application requested by a customer to scrape and classify text from open media sources like blogs and journals. Later some customer's requests for similar features showed up a new business opportunity. Due to the developed application's flexible design, the time required for the development process of a new company related product was shorted by 80%
Duties:
■ Design a text mining pipeline to classify text using a user-defined corpus
■ Develop a text classification module using the defined pipeline with SVM for the classification step. The data was processed using NLTK and Scikit-learn python libraries
□ Service Planner: Is a module for scheduling the medical guard service according to defined policies. This module increases the final product value by reducing the final user required time in the scheduling process, starting from an initial proposal that may require minor changes