Daniel Takabayashi
Job Position: Software Engineering Manager & Principal Software Architect
Location: San Francisco Bay Area - CA
I have created this page to show the most important projects I had the opportunity to work with, besides my technical certifications and some publications. If you need more information about me, send me an email at daniel.takabayashi@gmail.com.
Have fun!
Summary
Projects
Dr. Opinion
Comming Soon
Laguro.com
Comming soon
YelpSpy
Comming Soon
Apache Marvin AI Platform
Under Construction
Marvin is an open source project that focuses on empowering data science teams to deliver complex solutions supported by a high-scale, low-latency, language agnostic and standardized architecture platform, while simplifying the process of exploration and modeling.
Booska
Under Construction
Site to search for products on the web.
Role: Architect, technical leader and developer
Innovative project because it compares the price of other seekers (Eg. Buscapé and zoom) as well as being independent agreements with e-commerce, because all information is obtained from the web, from spiders programmed for this. The biggest challenge was to structure all the information obtained from the web.
The project is still in development but the first beta version was already available online.
keywords: natural language processing, computer vision, play framework, cassandra, python, javascript, java, high availability, python, elastic search
Scanboo
Under Construction
Platform for crawler, structuring and match (visual and textual) of products.
Role: Researcher, architect, technical leader and developer
First SaS platform in Latin America that allows the uptake of e-commerces data over the Internet to preprocessing and classification, aimed at monitoring of price positions and deduplication products on a scale from semi-autonomously. This project has given rise to two: Boomage and Booska. After nearly one year of scientific research, the biggest challenge was to overcome the scale in various fields (eg. Image processing, natural language processing, visual configuration of intelligent crawlers) with time constraints, besides the difficulty of gathering the skills needed to form the ideal team to make the product. In this project, several different architectures have been proven, many of the results caused frustration in the team, but at the end of 6-month project and a team of nearly 20 people (researchers, data scientists and programmers) the second beta version of the platform launched .
keywords: python, c language, java, rails, ruby, web crawler, natural language processing, computer vision, cassandra, postgree, django, celery, s3, elastic search, svm.
Boomage
Under Construction
Technology to recognise products from real photos.
Role: Researcher, architect, technical leader and developer
Innovative project that uses deep learning applied to computer vision to make it possible to search visually similar products on the basis of e-commerces. The product was engineered and the first stable version released and sold in less than 4 months. The biggest challenge was working with machine learning in scale (millions of classified and indexed images) to meet an online application (5s response time).
keywords: spark, graphlab, mahout, hbase, Cassandra, computer vision, caffe, python, aws services, gpu, c language, openmpi, opencv, OpenCL, distributed KNN.
Julia Architect
Under Construction
Intelligent system for recommendation of architectural solutions
Role: Researcher and developer
Julia Architect is a tool that, using AI techniques, is designed to assist software architects in architectural design decisions, reusing past solutions for today’s problems. Receives as input non-functional requirements and returns suggestions of architectural solutions.
Keywords: Case-based reasoning; Architectural tactics; Non-functional requirements; Quality attributes; Reuse of architectural decisions.
Oikos Hub Platform
Under Construction
Platform for distributing digital books (epub, pdf) comprises a DRM ecosystem (encrypted content distributor and mobile client for reading).
Role: Researcher, architect, technical leader and developer
Innovative project because it is an economical alternative to Apple and Adobe content distribution systems. The main objective was to provide small publishers the distribution of secure content in an economically viable way.
The beta version was created in just six months with a team of developers micromanagement. The challenge was to launch the beta in production with a minimum group of technicians without the proper knowledge in record time. The project was a success but was canceled for financial problems.
keywords: liferay, ios, android, phonegap, encryption, drm, java, javascript, objective-c, aws.
Publications
Marvin - From exploratory models to production
October 2017 Papis.io
Marvin is an open source project that focuses on empowering data science teams to deliver complex solutions supported by a high-scale, low-latency, language agnostic and standardized architecture platform, while simplifying the process of exploration and modeling. Build model-dependent applications in a robust way is not trivial, one is required to have knowledge in advanced areas of sciences like computing, statistics and math. Marvin aims to abstract the complexities in the creation process of a scalable, high available, interoperable and maintainable software.
BooViews: Aspect-based Sentiment Analysis on Product Reviews combining SVM and CRF in Portuguese
July 2016 PROPOR – International Conference on the Computational Processing of Portuguese
Customers use product reviews to gather opinion regarding a product before making a purchase decision. Reviews are available in several e-commerce businesses and wrote by real customers. Companies can use these reviews to harness consumer information by artificial intelligence algorithms and automatically extract product information, such as review polarity and product aspects. In this paper, we cover techniques to classify reviews polarity, extract product aspects and classify them. The resulting SVM classifier got 91.7% of precision in classifying the sentiment of the reviews, 74.2% of F1-score using CRF to extract the product aspects and 79.9% of precision classifying aspect’s polarity.
Preprocessing Images to Improve Deep Neural Networks Classification
October 2015 | Proceedings of XI Workshop de Visão Computacional
The amount of images used to train a classifier has a great impact in the algorithm’s performance. In the domain of e-commerce, most of the data set is formed by studio images and we demonstrate in this work that by multiplying the amount of images with different transformations can provide a significant boost on the overall performance of the classifier. Moreover, we list the challenges encountered in such task and present the improvements obtained with a classifier trained with additional images that were synthetically generated by applying several transformations on the original data set. Finally, we show that a Convolutional Neural Network, one of the current prominent algorithms in Computer Vision can be pooled together with binary classifiers for each class of images in order to predict images classes with great accuracy.
Certifications
Java Enterprise Edition 5 Enterprise Architect Certified Master (Part 1)
July 2013 - Oracle
IBM Certified System Administrator - WebSphere Application Server Network Deployment V8.0
March 2013 - IBM
IBM Certified Administrator for SOA Solutions - WebSphere Process Server V7.0
March 2012 - IBM
IBM Certified Deployment Professional - WebSphere Process Server V7.0
March 2012 - IBM
IBM Certified SOA Solution Designer
February 2012 - IBM
IBM Certified System Administrator - Connections 3.0
June 2011 - IBM
IBM Certified System Administrator - WebSphere Application Server Network Deployment V7.0
June 2011 - IBM
IBM Certified Application Developer - Rational Application Developer v7
January 2010 - IBM
IBM Certified SOA Associate (01/2010)
January 2010 - IBM
IBM Certified Associate Developer - WebSphere Studio, V5.0
November 2006 - IBM
SCWCD - Sun Certified Web Component Developer for the Java 2 Platform, Enterprise Edition 1.4
May 2006 Sun Microsystems
SCJA - Sun Certified Associate for the Java 2 Platform, Standard Edition, Version 1.0
June 2005 Sun Microsystems
SCJP - Sun Certified Programmer for the Java 2 Plataform 1.4
January 2005 Sun Microsystems