Verified Expert in Engineering
Data Engineer and Developer
Roberto is a software engineer with over a decade of experience primarily dedicated to building efficient data-processing pipelines in both hardware and software. He's worked in a range of industries/fields (media, NLP, defense, electronics) and tools/technologies (BigQuery, Python, VHDL, SQL, FPGA). Besides a demonstrated history of delivering solutions that vastly improve performance, Roberto brings an analytics-driven mind and a problem-solving attitude to any challenge.
Linux, Git, Jira, Visual Studio Code (VS Code), IntelliJ IDEA
The most amazing...
...thing I've (re-)designed was a GIS pipeline that became 10x faster and allowed new business cases to be feasible.
- Improved a core location data processing pipeline in both time (over 90%) and costs (70%) by redesigning it using Apache Beam, Google BigQuery, and Apache Airflow.
- Built internal tools and APIs using Angular and Flask to decrease the operational time used in audience segmentation setup by allowing account managers to handle customizations by themselves, saving 100s of developer hours per year.
- Led the back-end efforts to integrate existing marketing targeting and attribution products to create a better user experience by designing a new internal API that made negligible the discrepancy seen by clients in reports.
- Improved several pipelines orchestrated by Apache Airflow by optimizing queries in BigQuery and Redshift, Apache Beam dataflows, and the overall architecture and data models.
- Led the development of a new product used for building custom audiences by self-serve clients with a multi-cloud (AWS/GCS) solution—extending the existing internal solution to support further customizations and built an entire API surrounding it.
Machine Learning Developer
- Designed a real-time sentiment analysis service using Flask, MongoDB, NLTK, and integrated IBM Watson speech to text API for use during meetings on Zoom, WebEx, dial-in platforms and others.
- Built a RESTful API around Kaldi (open-source library) to provide a speech-to-text service optimized for long conversations between 2+ people in noisy environments as an alternative to the existing cloud options using Python WebSockets library.
- Developed tools to evaluate the performance of core API components using Python and Node.js.
- Fine-tuned speech recognition models to increase the accuracy of transcriptions for enterprise clients.
Software Engineer, Graduate Research
Western University Canada
- Designed a framework for predicting the users' current activity spatial preferences using Python.
- Preprocessed location datasets with hundreds of millions of samples to prepare them for the above prediction task using Python.
- Produced a dashboard summarizing over 30 million real-time auctions to help the decision-making process of parameters used in web advertisement inventory.
- Provided support and guidance to students in the database course as a teaching assistant.
Field Application Engineer
- Supported business development and demand creation for electronic parts in more than 20 customers by providing technical advice and discussing the full lifecycle of products.
- Served as an FPGA specialist for Arrow Electronics in Brazil and increased the adoption of the technologies by presenting seminars and providing in-company training.
- Managed supplier and customer expectations by building and developing. relationships and keeping stakeholders up to date on activities and plans.
- Supported field sales in their pre- and post-sales activities by providing technical support and consulting (based on knowledge of products related to the company line card).
- Designed a protocol with forward error correction used for communication in a high-noise environment (near high-power RF elements) and implemented it in VHDL.
- Automated the process of generating KPIs by building tools to gather data from SVN and Redmine and built a visual dashboard using Python.
- Reduced the time spent to setup the development environment by creating scripts that automated the task using Bash.
- Implemented an interface for the DDR2 controller, which reduced over 90% the number of lines coded needed to use the controller (VHDL).
- Added code coverage to the design flow by increasing the test coverage to over 10% average when compared to the existing tests (prior to additional code coverage) for VHDL.
- Ported signal processing algorithms used in remote sensing, defense, and surveillance radars; designed it in IDL to hardware while using VHDL. This includes optimizing the original design to better suit a hardware implementation targeting FPGAs.
- Designed the FPGA architecture of an airborne remote sensing radar and led the FPGA team working on the project.
- Improved the training resources and trained and mentored junior engineers.
- Designed a library to facilitate the connection of any peripheral to the processor local bus (PLB) using VHDL.
Digital and Embedded Systems Developer
Idea! Electronic Systems
- Ported digital signal processing algorithms used in real-time embedded applications (TV broadcasting equipment for ISDB-TB standards) from MATLAB to VHDL to run them in FPGAs.
- Created the verification library used internally in all the companies' projects using VHDL and C. Library was composed of data generators, validators, and transformers used in the verification flow.
- Reduced the time spent to set up the development environment for new modules by over 80% by creating scripts for automating the task using Bash; also maintained the scripts (Bash and Makefiles).
- Developed a proof-of-concept GUI for displaying television channels using Delphi in less than a month without prior knowledge of the language.
- Built an API that provided a simple interface for third-party developers to implement interfaces for a digital TV USB dongle using C.
- Developed CLI tools for test automation using C and Python which significantly reduced the test setup time; also automated some of the recurrent tests for a new board.
Users' Current Activity Preferences
Python, SQL, VHDL, C, Common Lisp (CL), Prolog, Java, Bash
BigQuery, Apache Airflow, Apache Beam, Emacs, Altera Quartus, Xilinx Ise, Radar, Kaldi, Git, Jira, IntelliJ IDEA
Data Engineering, Google BigQuery, Data Processing, Big Data, Machine Learning, Supervised Learning, Compilers, FPGA, Data Structures, Algorithms, Networks, Operating Systems, Distributed Systems, File Systems, Image Processing, Evolutionary Algorithms, OOP Designs, Digital Electronics, Artificial Intelligence (AI), ASIC, Digital TV, Digital Signal Processing, Device Drivers, Radar Remote Sensing, Systems Engineering, Embedded Hardware, Business Development, Network Processors, Natural Language Processing (NLP), Speech Recognition, IBM Watson Speech-to-Text, Digital Advertising, Geolocation, Data Warehousing, Data Warehouse Design, FastAPI, GPT, Generative Pre-trained Transformers (GPT)
ETL, Database Design
Databases, Relational Databases, MongoDB, Amazon DynamoDB, Redshift
Pandas, Scikit-learn, XGBoost, NumPy, Node.js, Natural Language Toolkit (NLTK)
Unix, Linux, Amazon Web Services (AWS), AWS Elastic Beanstalk, Docker, Visual Studio Code (VS Code)
Master's Degree in Software Engineering
Western University - London, ON, Canada
Bachelor's Degree in Computer Engineering
University of Campinas - Campinas, SP, Brazil