Search for "AI Engineer" Talent for Indonesian Startups

Have you ever heard the term automation, chatbot, to artificial intelligence? This technology used to be considered a miracle. Now the technology continues to be studied and developed by engineer AI (Artificial intelligence). They are the people behind the robots that reply to our chats quickly, systems that are able to calculate and visualize data, and the like.

DailySocial trying to explore how startups that focus on AI services or have a dedicated AI team or division develop teams and grow each other's knowledge and skills.

100% local talent

Volantis, for example, is a company that offers data management services using AI and Machine Learning technologies, both for prescriptive and predictive purposes. Now they have 12 a software engineer and 6 data scientist. All of them are local employees who have been trained since 2016.

"We believe that seniority is not everything and we believe that bright children if given the training and opportunities will very quickly acquire skill. Most of the engineer we are a cadre of fresh graduates who we taught from the beginning. We also believe that problem solving, basic science, logic and math strong is very instrumental in quality engineer. Most of the engineer are pure (science) children, whether from physics, mathematics, or computer science," explained CEO Volantis Bakhtiar Rifai.

One area of ​​AI that is widely applied in Indonesian businesses is NLP. Using this branch of science, technology is able to read Indonesian input and then respond naturally. Everything is automatic. AI technology plays a role in translating intent and providing relevant responses.

The startup that applies this technology is Kata.ai. Kata.ai CTO Pria Purnama explained that currently they have 2 Senior Research Scientists and 1 Machine Learning Engineer. All Indonesians. It's just that they are graduates from abroad, namely Italy, England, and Germany.

"There is one common thread: Indonesians who understand the challenges of making Indonesian NLP," he explained.

In almost the same line of business, there is Prosa.ai. Teguh Budiarto, CEO of Prosa.ai, told his side today build building blocks technology in the form of AI functions to generate analysis and the ability to respond to data. The results are implemented in the form of a chatbot, sentiment analysisand regulatory technology for text data. Voice data will come up with a solution voice biometrics, call center transcription, etc; while the image and video data for people and vehicle management system.

"We have more than 30 AI people engineer and more than 20 people software engineers. Assisted with data preparation by more than 60 annotators, be it for text data, sound and image or videos. 100% of them are local talents," said Teguh.

Meanwhile, Bahasa.ai combines AI and WhatsApp capabilities for a service that automatically accommodates user needs. They have a number of well-known clients, such as DANA, Smartfren, Bank Sinarmas, and Sociolla.

"The core team that developed core modelsnatural language processing we have 3 AI engineers (including one founder), some help a software engineer which helps implement the model in real business use. Until now, we are still assisted by 100% local talent. Since our business is developing AI technology for understanding Indonesian, we want all our teams to have a complete language context—which only native speakers of the language can get," said Co-Founder & Chief AI. language.ai Samsul Rahmani.

Then there is Nodeflux. Startups engaged in the participants' vision This AI provides a solution to convert image or video data into structured data or useful information.

Nodeflux co-founder and CEO Meidy Fitranto explained, currently they have 50 engineer and all local talent.

"Yes for talent engineer AI, for our cadre there will be some kind of mentoring from senior level AI engineer ke junior level, where seniumrAI engineer will be together junior engineer complete computer vision case, starting from making analytics model to optimization and experiment"said Meidy.

DailySocial also spoke with two Indonesian unicorn startups, Bukalapak and Gojek.

Head of AI Research Bukalapak Muhammad Ghifary said that they have a special AI division. The division focuses on providing and innovating AI-based products and services. There are two processes related to AI, namely R&D and engineering. The R&D process aims to produce an AI model or solution that performs certain functions. While process engineering carried out so that the AI ​​solution can be implemented and used on a large scale.

"Currently there are approximately 14" engineer AI, all of which are domestic talents with educational backgrounds at home and abroad," explained Ghifary.

Di Gojek, the team responsible for developing and implementing AI technology is a team data science. The team follows the CRISP-DM (Cross Industry Standard Process for Data Mining) process. Gojek have about 40 data scientist and 20 machine learning engineer.

"As global company"Our team is spread across various countries, namely Indonesia, Singapore, India and Thailand," said Senior Data Scientist Gojek Ardya Dipta.

Team management and development

At Volantis, the formation of the AI ​​team went through several screening stages, both through written tests and interviews. Talents to look for are those with logical abilities and problem solving brilliant. When they have successfully joined, they will be supported by qualified research and development facilities so that they can develop and provide services value to the company.

For Kata.ai, they are active and routinely perform surgery paper, new innovation whatever has been published will be discussed. If something is judged according to need, they will develop it further.

"Results may change improvement on the platform kata.ai and also we republish. Usually we publish 2/3 papers per year at conferences such as INACL, ACL, CICling, and the like," explained Pria.

Development and review papers also done by the team engineer AI in Bahasa.ai. Each new AI-themed research will be discussed, discussed, and implemented if necessary. This they do to stay Update to technological developments.

Management and development at Prosa.ai begins with a selection with fairly complete qualifications. Within the company they conduct cadre by organizing training which is divided into several levels, sharing development of AI technology, to sending team members to attend trainings, meetups, and similar events at home and abroad.

"Prosa provides infrastructure facilities, as well as support for training costs according to recommendations from supervisors / leads in each division, which can be submitted by each engineer. If the proposal is approved, they can follow it with full or partial support," added Teguh.

Meanwhile at Nodeflux, team engineer AI must at least have an interest and ability in a related field. Furthermore, when they become part of the team they will get a session peer tutorials, fellow place engineer AI shares knowledge.

"Reference sources for AI-related ideas are usually from research paper sourced from leading institutions, journals and conferences. Even then not all can be applied. Engineer AI regularly discusses research paper and its applications and create experiments that can be applied on the job," explained Meidy.

Meanwhile, to become an AI engineer team at Bukalapak, several skills are needed, such as data processing and analysis, developing AI models/solutions, to implementing AI-based software on a large scale. In addition, it also requires mathematical and problem solving the strong one.

According to Ghifary, Bukalapak provides an opportunity for team members engineer AI to evolve through development programme they have. Training and knowledge sharing It's also a tradition within the team.

Ardya tells environment di Gojek very supportive for the development of knowledge and skills. Besides the sessions knowledge sharing routine they also have study groups which every week discusses the blessed books.

"The transmission of knowledge can also be done when working on a joint project, namely by data scientist the more senior one gives mentoring to a more junior while doing review periodically in depth. Every review well documented and available sign-off from the mentor that the projects carried out are in accordance with the standards," added Ardya.

Is it hard to find talent?

The growing number of companies delivering and developing AI solutions creates a need engineer Increasingly AI. Bachtiar assesses that in Indonesia there are many great talents, but the main difficulty encountered is forming a culture of initiative and strategic thinking.

"Our weakness is sometimes we tend to be passive even though we are actually capable," he explained.

Meanwhile, for Meidy, finding talent in general is a challenge. The focus is not only on skill but also attitude appropriate to the company's culture and ability to collaborate. Moreover, AI is an area that is still growing.

"It's easy to be difficult. Of course experienced talents more difficult because apart from AI companies in Indonesia there are not many, engineer AI tends to be interested in text data or speech or participants' vision only [one branch of knowledge only -RED]. It's not easy to find one with good performance," said Teguh.

Bahasa.ai itself focuses on finding talents who have advantages in two fields, because there is not much special education in the field of AI studies.

"Usually we search between 2 kinds of people: graduates computer science who are proficient in mathematics, or graduates of mathematics who are proficient in Programming," added Samsul.

While Gojek sees that there is still a fairly high gap between the need for and the availability of talent engineer AI in Indonesia. Gojek trying to present several solutions related to this by presenting GoAcademy, Upscale, Rise, Gojek Xcelerate, up to Community up and meetup data science.

"Currently Gojek has thousands of employees in Southeast Asia with more than 30 nationalities. We hope that the assimilation of these talents can encourage acceleration for development skill and produce world-class talents, as well as gain exposure to multinational companies with various challenges and key learninghim," concluded Ardya.