Conversations with Lee Jin

I am a Project Manager here at Taiger, and I deal with project delivery. When we partner up with clients, we engage with them to identify the problems that they face and brainstorm possible solutions to tackle them.

Taiger looks into ways that AI can be integrated to enhance our client’s processes. Basically, you get to optimize your processes, improve certain things, or even realign some of the resources to get things done better. Our products comprise of chatbots, search functions, and the extraction of information from documents, which mostly used to be very dependent on human beings. Today, Taiger is leveraging on technology to improve these processes with automation.

As a Project Manager, I with the team to ensure we deliver the products on time to our customers. In Taiger, we consist of different groups that handle various tasks. These include front-end developers who build the UI and UX, back-end developers who deal with software, and a group of people handling AI and Natural Language Processing. We bring together these three groups of individuals and we generate solutions and customisations for our clients.

Here at Taiger, we are dealing with cutting-edge technology. We are talking about things that are at the front of the curve and we look at problems in a forward thinking manner – no longer are we talking about conventional systems. Since we are dealing with AI, we have to consider how things can go wrong with machines. AI should only be brought into processes or businesses that are able to understand and accommodate all these errors and implement AI. Our solutions support human employees in their work and supplement their workflow.

Yes, we are implementing IT solutions like content management systems. AI, however, can achieve much more, even replacing someone speaking on a phone. You can replace someone carrying out searches or someone who only conducts email routing from department to department. This certainly means replacing jobs that used to be manned by someone. In a way, the technology that we deal with is a lot more forward thinking. The technology that we use – deep learning, machine learning – is a lot more advanced. A lot more convincing is required when we go to clients to teach and explain to them our products.

For example, you may be an officer sitting in an agency. You need to deal with emails that come in regularly. In 2015, you probably had 200 emails. You will have one person dealing with 200 emails per month. In 2016, you have that doubled to 400. Do you require two people or do you want to advance the technology to route the emails automatically? Do you pay to streamline these processes moving into the future with the exponential growth of emails? Or do you continuously add more operations personnel to handle the load?

That is the kind of mentality we adopt when companies come to look at AI solutions. How do you replace more of the manual bit, so that money can be placed somewhere else, for instance  the sales side, etc.

For me, it tends to be very uncertain – days can be packed. At the moment, I’m handling roughly 12 different projects from various agencies, government entities and private institutions, all of which have varying timelines. Imagine that within a month, you have activities across all 12 parties with different stakeholders that will come to me for updates and discussions. Day by day, there are ad-hoc tasks, meetings, and that’s in addition to catching up with the team on how they are doing on a certain development for a particular project. At the same time, being in a start-up, we work with many different teams. Even though I am a project manager here, I sometimes have to dwell into the HR bit, and I also handle the internship programme at Taiger. For the internship programme, getting people in, recruiting, advertising, going for talks and panel discussions – there tends to be quite a lot of things to handle every day.

Yes, it is definitely hectic. You will have days where it is really packed and days where it is less packed. It is pretty sporadic and random, which keeps things interesting.

I used to work at a bank. However, being in a start-up, you get to see how the company grows. When I first started, It was a really small team. Roughly about 4 people in the development team. Right now, it is more than 20. You grow together with the company. You grow from series funding to series funding. Every single effort that is put in will be contributing to the growth of the company. That is the thing that I look forward to the most. Every small action that you make, generates value somehow, someway. Decisions that you make determine whether your customers come back to give you your next set of revenue. Every action by anyone here is in fact contributing to the company’s growth. Not just that, I think that because we are smaller, we are a lot more agile. You tend not to see your colleagues as colleagues, but as friends. Friends that you work with every other day to solve problems together, which is the fun part.

Yes. Start-up culture is usually managed by objectives. In a sense, you can come in at 9am or 10am, and you could work from home, as long as you deliver what you are tasked to do. At the end of the day, we look at your contributions and your value to the team. You are hence determined for your performance review based on that. To us, turning up or not is fine, as long as you are responsible for your own task, but that also depends on which company you are working for.

I think the difficulties are aligned to what I mentioned earlier. The reason is that as we grow, we are always talking about how to get from point A to B, and ensure that we deliver projects that customers are satisfied with. Difficulties that arise are always tied to that. For instance, not having enough people in the team, or clients being unhappy with the product. It’s not uncommon to have to put on multiple hats depending on what is required. We improve and learn from our mistakes and we move on. Every day, these are problems that arise in Taiger.

To get technical, there is this thing called precision and recall. Precision is how accurate your responses are. Recall refers to the kinds of search results that you are collating as correct. These are definitely technical issues that are linked to AI. We should always have expectations as to what we are aiming for; when we go to a client, we see if they want their product to be more precise or have a wider range of results in exchange for compromised precision. You can therefore set a target and move from there. Expectations management is key here.

There are clients who tend to expect something that can handle everything given a limited budget. However, that is not really possible as there are bound to be constraints tied down to the project or due to limiting capacities of the technology that we are equipped with. Hence, we have to manage that too. It is up to the Sales department to manage that alongside the people working in Product Delivery.

It is a long story, so I will just share the interesting bit. When I first started out, I was in my university days, after NS. I was interning at one of the government organisations. When I was carrying out my internship, I was dealing with emails that were coming in. For government organisations, you had to reply within 7 days. The organisation I was working for was called Service Excellence – to see how fast how complaints are taken care of. My boss asked me to look at a set of data and asked how I can optimize the data. At that point, data analytics was not very fancy, since it was just up and coming, similar to AI. Then, I looked at the data and generated a list called elapsed time – how long an officer takes to reply a particular email. There were a few ‘black sheep’ that were pointed out, and when I handed this information to my boss, he checked on the information and realised that the team lead of the team says that they are not performing and were due for a review. Because of the search, they did not get certain incentives to their job.

During the internship, I learnt about data analytics with the resources that my university provided me with. Then, I moved on to an exchange program in Germany. At that point, I was too engrossed in data analytics, so I took up a masters course in AI and web data integration and web mining to really focus on the data and AI aspects. At that point, my resume looked great. Usually, courses you take on exchanges are pass/fail courses. But I thought to myself, why not take up a few masters courses and have them under your belt? That was my mentality then.

Once I honed my skills in Germany, I started applying for jobs. My first job was at UBS and that was where I started on Ontology. I was given a job of Ontology within the bank and I had to perform. Banks are very competitive – out of 9 interns, I was the only one who was given a full-time position. But my manager was telling me that doing the tech work may not be my thing; I remember him saying “you are much better at management. You should place your money where your mouth is and do something in between.”

When I came back after graduation as a project manager, I was thrown at the deep end of the pool. I was handling construction projects, which was totally different from software engineering. Then, I went to Hong Kong and switched my job from hardware to software. In Hong Kong, I met another mentor who told me to get out of hardware and move to software. Then, I saw Taiger’s advertisement, which seemed to match what I wanted and thereafter, I applied. After speaking to the CEO, I realised that I should make the switch and get out of somewhere structured and try something where my passion is, which is how I ended up at Taiger.

Definitely, you need to hedge against a risk. You need to know that your risk should be covered by certain rewards. In a start-up culture, the reward would come in terms of equity. That means that I am betting myself with the company that I can perform. If I can perform well, the company performs well and my equity will grow. At that point, my decision was made because of the prospects that I had if I shifted. I talked to a few people who did the same thing, jumping from a structured organisation to an unstructured organisation. It also depends on the character of the person. For me, there was nothing much for me to consider. At that point, I was just 27 and there was nothing for me to lose. Till now, I think that it was a good decision. When you do things well and perform, it is very easy for the people in the company to know. How do we validate that? Instead of climbing the corporate ladder, you can tell this from performance reviews and the sincerity of the packages that they give you. This can be seen when I brought up the internship program for Taiger. Back then, there wasn’t such a program. But there was an opportunity and I felt, why not.

I probably would not have gone for hardware, and would have honed my skills more on the software aspect back at UBS. But that was the first opportunity and I couldn’t say ‘no’. So, in a way, it was still a fulfilling opportunity. I would not have changed much, other than to do better at JC – I didn’t do well in my JC days.

Army also played a part in changing my character. I was in Commandos back then and the mentality that I developed there gave me the motivation to pursue my goals. When I was in secondary school, I was in the Art Elective Program. Then, I went to the science stream and flopped. But army changed things. Being in Commando’s really changes your perspective on life. Many things stuck with me after having gone through army. For instance, near death experiences will always stick with you. Your perspective on life definitely changes. When I went into IS, I had the mentality. Since I am already here, why not go all out?

Yes. I went to many different courses, such as becoming a certified ethical hacker. Right after National Service, I had roughly 9 months to myself. During that 9 months, I went to volunteer and was clearing my community service requirements for university. I also went for forensic investigation courses, like the Cisco Certified Network Associate (CCNA) which is the network-related certification, to prepare for university. When I got into SMU, I already cleared that portion and thereafter, I could move on to doing other things such as organising events for GRC, meeting Ministers like PM Lee, etc. That 9 months is a lot of time where you can prepare for university.

That would be from my army experience. It was during one of my exercises where I had a near death experience where I fell from a cliff, and would have died if not for my field pack, which cushioned by head from the impact. That forged my perspective on life. That experience forged the mentality in me to really maximise my days for the immediate next phase of my life as well as to impact others through volunteering.

Another would definitely be my first internship experience, as well as being able to meet the mentors who guided me along the way throughout my life.

There was another other experience I had when I first joined Taiger. During that period, I had just switched from a structured working environment to a start-up. I tried to implement what I knew in Taiger and brought it up to my CEO. When I did, we had a discussion and he told me, “we are problem solvers here. Let’s see how we can take these constraints and move beyond that. Unlearn everything and see how we can work around things. Look beyond your job description. Do what is good for the company from this point.”

This teaching stuck me at the right time as I was transiting and at the same time, I wanted to see how much value I can create at that juncture. Given the past year, I learnt a lot from his advice.

AI is still a new thing. People need to be comfortable with AI before they can take up AI solutions. There is a lot of hearsay, such as “AI is so powerful. AlphaGo for example, can beat the top chess player.” In Singapore, there are a lot of initiatives emphasizing AI and automation. I believe that it will be very impactful at changing many industries, especially those that were untouchable in the past such as accounting and law. For example, a 200 page law document can be analyzed with the help of an AI. That is definitely a game changer to the industry. It will not put lawyers out of their jobs, but allow them to concentrate on other things.

We are very far away from automating the world. If you look at AI, we are always dealing with a small portion of your brain – the cognitive bit. Or, for example, robotics. You are only simulating the hand to drill things, or to understand language. All these are small feats as compared to what humans can achieve. But to combine everything is really farfetched. For AI’s to be fully automated, we will have to wait a few decades.

Definitely positive, or I wouldn’t be working in AI. We do not try to see it as the replacement of jobs or the taking of people’s livelihood. It is not that case. It is more about if we can do so much more, why not. There are so many things we can deal with instead, like solving environmental issues, eradicating poverty, etc. If we look more into that, maybe our resources might not run out so fast.

First is definitely the mindset. People are concerned about AI taking over their jobs. Secondly, it is the effectiveness of AI at its job. There are many things that can go wrong with regards to AI. Look at the Microsoft Chatbot for instance. If a Chatbot were to be implemented in say, a government organisation, is the AI going to run smoothly as it is supposed to, or cause havoc?

If we look at hard skills, I would say, being able to program and understanding programming languages like Java, Python etc. That would help you in understanding how things work. Other things include machine learning and deep learning. You have to keep yourself up to date with the technology out there.

For soft skills, I would say being able to manage events and being able to deal with people and problems.

Both hard and soft skills are equally important.

We definitely need to invest in our developers. Engineers need to understand data science, and be willing to explore and break boundaries technologically-wise. We need engineers who are able to apply their knowledge into AI.

Where you are right now, try to talk to people and evaluate all options available.