Having been in the IT industry for 17 years, the company decided that they would share their learnings with the next generation. And so, on a typical Thursday morning, approximately 150 students from universities across Sri Lanka made their way to Trace Expert City. They were here to find out what CodeGen was creating behind its doors. They were here for Ennoble 2017.
The opening keynote
The keynote speaker of Ennoble 2017 was Dr. Harsha Subasinghe – CEO of CodeGen. He opened by sharing that the biggest contributors to the Sri Lankan economy are the unskilled workers, the majority of whom are working in Middle Eastern countries.
CodeGen wants to change this and to make Sri Lanka a techno-based economy that runs on delivering higher performance solutions to the world. But the biggest challenge to achieve this vision is talent. Sri Lanka has highly skilled talent, but few people that can build products that add value to people’s lives. He then compared the current Sri Lankan debt to the value of Uber.
Dr. Harsha then said, “If we had a single company like Uber then we can pay $35 billion to pay off the debt and another $35 billion to build the infrastructure of a better country. So the question is can we build companies like Uber and Airbnb? All of you guys in the technology industry are capable of doing so.”
He then shared that in the past, Sri Lanka built engineering marvels like Sigiriya. But after eras of colonization, Dr. Harsha argued that we adopted a mindset of slavery. He went on to say that it still exists today with many university students.
Dr. Harsha then went on to share the origins of TRACE. They wanted to reawaken Sri Lanka’s culture for excellence and reverse the brain drain. He shared that free education isn’t free because it’s paid for by taxpayers. So when a university graduate leaves for greener pastures, it costs the country millions.
But he admits that we also need an ecosystem to create great products. Thus, they pitched the concept of TRACE to the government and TRACE Expert City was born. Click here to know more about the origins of TRACE Expert City.
Afterwards, Dr. Harsha shared how CodeGen was born and is now a large company with over 400+ employees. It’s also won countless awards and presented itself at various conferences. He shared a few examples of CodeGen’s clients before delving into the technology stack used to support these clients.
He commented that CodeGen was one of the first companies to adopt a service-oriented architecture. After sharing some of the avenues they’re exploring with AI, Dr. Harsha shared the architecture of TravelBox, which is one of the company’s premier products.
Dr. Harsha then shared the progress of the VEGA electric supercar. It’s been almost 3 years since they first started the project. They’ve faced quite a few challenges as they’ve had to build the car from scratch. Nonetheless, they’re aiming for the car to go from 0 to 100km/h in 2.9 seconds, a top speed of 240km/h, and a range of 240km. The company is aiming to release a prototype by December this year.
Additionally, they’re also hoping to reduce its weight and make it capable of autonomous driving. But the supercar isn’t the only electric car project the company is pursuing. The company is also pursuing another project to create Mini Coopers with electric engines.
But this isn’t the only moonshot project CodeGen is exploring. Another project is AiGROW, which aims to use the power of AI and robotics in hydroponics to improve farming. There’s also a project they’re pursuing with the University of Peradeniya to build advanced drones. And of course, there’s the XOLO Smart Classroom. Supporting all this is the QBITS Labs that explore the possibility of introducing AI into a variety of fields.
“Always help people. If do good deeds then good things follow.” – Dr. Harsha Subasinghe
Ultimately, CodeGen aims to become the first billion-dollar company from Sri Lanka. And this requires an innovative mindset. But the most important thing according to Dr. Harsha is that you ensure that the people you work with will grow over time. He concluded by encouraging the participants to do good deeds and also to use technology. Use technology to make something for Sri Lanka and help the people of the country. Help the people by fixing the economy with technology and create an advanced country.
Improving app performance
Following Dr. Harsha’s keynote, we saw Tharanga Hewavithana take the stage. He opened his session by sharing that a response time of 100ms gives users a feeling of instant response. A response time of 1 second keeps the user’s flow of thought seamless. At a response time of 10 seconds, you’ll still have the user’s attention but any more and you’ll lose it.
This is why performance is so important. Tharanga then shared that 90% of the factors that affect this are found in the front-end of your application. However, when it comes to performance, response time is only one factor to consider. He shared that there are other factors such as scalability, elasticity, and downtime that should also be considered when gauging the performance of applications.
Tharanga then shared the story of the TravelBox platform, which was built by CodeGen. During his presentation, he shared how the architecture had evolved over the years. He concluded by sharing the advantages of adopting a distributed architecture for apps. The advantages are better performance, more elasticity, and higher availability.
Practical AI for Consumer Software
“What does AI mean to you?” said Yohan Welikala as he opened the second parallel session. He then went on to introducing what Artificial Intelligence is. But if you are to truly understand AI then you have to understand abstract concepts like intelligence itself. He then shared that CodeGen is working towards helping computers better understand humans and carry out tasks.
To give the audience an idea of what kind of work they were doing, Yohan shared a picture of an article. This was an article about quantum computing. He then asked the audience, “Can you read this article?” The article was in English so we understood that it was talking about quantum computing.
He then shared a picture of the same article in Greek. Nothing made sense and Yohan said, “This is what the computer sees. To a computer, the original article makes no sense.” He then introduced us to ontologies, which are being used to help computers better understand languages. Yohan concluded by sharing what CodeGen’s newest recruits were working on.
Afterwards, we saw Tharindu Ranasinghe take the stage to continue the session. He shared that another challenge CodeGen is working on is to make AI as friendly and as interactive as a human travel assistant would be. To do this, they aim to utilize natural language processing. Tharindu went on to share that for this information is vital and they’ve built tools for this as well. Once the information has been collected, it’s fed into an algorithm that recommends locations for a user to go on a vacation. He concluded his session by sharing the CodeGen AI technology map.
The Ennoble 2017 tour of CodeGen
Following the parallel sessions and lunch, we were then taken on a tour of CodeGen. The first part of the tour shared how the company builds its software and the different roles involved. Afterwards, we saw another session recapping the different research projects the company is involved in, which includes Artificial Intelligence.
Once the session ended, it was time for the second half of the tour. This half focused on introducing the different projects CodeGen is pursuing. The first was the VEGA electric supercar. We found that much of the technologies that power the VEGA are being built in-house. This includes a powerful drivetrain that the team is building, which will help it go from 0-100 km/h in 2.9 seconds. To meet their design goals, at the moment the team does much of their testing using computer simulations.
The second project we saw was AiGROW. This is an agriculture related project that harness the power of Artificial Intelligence. The goal of this project is to create a safe and sustainable food supply for the future. As such, it aims to automate the agricultural process and reduce costs. At the same time, it also focuses on enabling urban agricultural applications with advanced technologies.
To do so, AiGROW utilizes vertical farming in a computer-controlled environment. Inside this environment, every factor from temperature to humidity is controlled to help maximize crop yield. Afterwards, we got a glimpse of VEGA’s electric mini cooper car. This project takes a classic mini cooper and aims to enhance the performance by converting it into an electric car.
We then got a closer look at how CodeGen is utilizing AI to build autonomous driving solutions and monitor the health of plants. And the final stop on the tour was CodeGen’s drone project.
The project done in collaboration with the University of Peradeniya aims to create advanced algorithms. These algorithms would, in turn, help almost any type of drone to navigate obstacles while hovering stably. In a demonstration, Dr. Harsha encouraged participants to crash the drone into a tree. However, every time the drone detected and avoided the tree on its own.
The final panel discussion
With the conclusion of the tour, everyone returned to the Trace Expert City auditorium. This was where the final panel discussion would take place. One of the first questions thrown at the panel was regarding project management, which was answered by Yohan Tsung,.
Yohan answered this question by sharing that as a project manager, “It’s your responsibility to visualize on a future day what’s going to happen and then make a decision.” As a project manager, you must foresee and future issues a project may face. Once you have this information you prepare a plan for 3 – 6 months to achieve the due target. But if you want to truly succeed as a project manager then you must also learn emotional intelligence.
Of course, a few questions were thrown at the panel regarding the VEGA supercar, which were answered by Dr. Beshan Kulapala. One of them was regarding the carbon fiber body, which Dr. Bheshan shared is only being used for the exterior panel. He also shared that the most expensive and fragile component of the car is the batteries. They are installed from the bottom and can be removed by raising the car.
During the panel discussion, we also saw Tharindu Ranasinghe share some of the tools CodeGen uses for its AI efforts. He shared the Apache Spark is one of the open source tools they use. But it doesn’t cover much about machine learning algorithms, which is where Python comes into the picture. But for deep learning, Tharindu recommended having a look at Tensor Flow, Keras, and Café. He went on to recommend R if anyone was interested in carrying out data analysis.
With the conclusion of the panel discussion, Ennoble 2017 came to an end.