The Women in Tech Show: A Technical Podcast

Data is constantly flowing through a system. From telemetry data to data from IoT devices, as the amount of data increases, challenges emerge on how to process it. Holden Karau, Open Source Big Data Developer Advocate at Google, talked about how data can be processed and analyzed in batches or in streams. Holden explained the infrastructure needed to process data at scale and fundamental performance improvements for these systems.

When you’re working on a product, an important part of the process is to understand how your customers use it. This helps drive improvements. Avni Patel, CEO and Founder of Poppy explained the role of consumer psychology in analyzing products. Avni also explained the process of launching a tech startup as a solo founder. We talked about how she built and launched the first version without having a background in software development and programming. Avni explained her experience at YC and how she handled rejection the first time she applied.

A data-driven approach is becoming more common across companies from different sectors. These companies have people that are focused on getting insights from data, such as data analysts and data scientists. Hanna Landrus, Data Scientist at Usermind explains the difference between data analyst and data scientist. Hanna has experience in both these roles. She explained the types of problems that she explored as a data analyst and as a data scientist. We also talked about the mathematical concepts present in this area, and the workflow and technologies she uses.

As the amount of data has increased, we have been able to apply other machine learning methods like deep learning. Erika Menezes, Software Engineer at Microsoft explains what deep learning is and the problems it has been able to tackle. We talked about artificial neural networks and how these are the core component in deep learning architectures. Erika also explained how a deep learning project is structured and the different steps in the pipeline.

What does it mean for a machine to be creative? Maya Ackerman, CEO of WaveAI, addresses this question. We talked about the challenges of building systems that have creativity. Maya explained how she began exploring the idea of computational creativity by building a system that makes songwriting accessible to more people. This system, called ALYSIA, plays the role of a creative partner and helps a person write music. We talked about how the system was built using machine learning and the ways in which it is creative.

The workflow of a software developer has evolved a lot throughout the decades. With these advancements we have become more productive and focus on building things that are core to the business that we work on. Simina Pasat, Senior Program Manager at Microsoft, explains what developer productivity consists of. We talked about the developing pipeline for building mobile apps and the components that can be automated. Simina also explained the bottlenecks in developing a mobile app and how to monitor apps and learn from data.

Technology is a broad field that touches many other fields. Camille Eddy, Mechanical Engineer explained different areas in technology that she has worked on. We begin the discussion with machine learning. Camille explained different machine learning systems we interact with, how bias can be present in these systems, and ways to combat it. We then talked about Mechanical Engineering and discussed the process of building a robotic hand. Camille explained the design and prototype process. We also talked about the hardware and software engineering components of this process.

When we open Netflix we get pretty good recommendations for content that we may like. This is one example that shows the data-driven mentality at Netflix. Julie Pitt, Director of Machine Learning Infrastructure at Netflix, explained other ways in which Netflix is using data and machine learning. We talked about using data for creating content, and for changes in the infrastructure. Julie explained how her team is working on improving the workflow of data scientists. We talked about the bottlenecks and the major interface between data scientists and a Machine Learning platform.

Web and mobile applications can become more popular than we had anticipated. If we’re not prepared, our application can have downtime. If we are prepared, it means that our application can scale. Scaling an application comes with numerous challenges. One of those is how we reason about system performance and how that influences architecture changes we make to our system. Kay Ousterhout, Software Engineer at LightStep explains what performance means and why users struggle to reason about it in today’s systems. Kay discussed her work on Apache Spark and how jobs were decomposed to provide clarity about performance to users. We also talked about detecting bottlenecks in a system.

Cybersecurity has been around for several decades. With technological advances also come more advanced attacks. Michele Guel, Distinguished Engineer and IoT Security Strategist at Cisco, has been in cybersecurity 20 years. She explained how the panorama has changed, the different types of attacks, and also what the security threats in IoT are. In this episode we were also joined by Mandy Galante, Cybersecurity Educator at Red Bank Regional High School. Mandy has been teaching cybersecurity since 2002. She explained how teaching cybersecurity has changed, the tools they use, and training competitions that kids can participate in. At the end we talked about Cyberstart, a program to educate young people about security through games and challenges.