How a Microsoft Software Engineer Evolved His Career with Codesmith

It's no surprise that understanding AI is becoming a key component of the software engineering career. But how and where can software engineers learn the right AI tools and skills that will advance their careers? We caught up with Will Kencel, a Software Engineer (SDE II) at Microsoft, to learn how Codesmith’s AI-ML Technical Leadership Program gave him hands-on experience with AI tools. Find out what he built for a final project with a few teammates, and his advice for software engineers looking to get a leg-up in this quickly evolving industry. 

What prompted you to enroll in Codesmith's AI & ML Technical Leadership Program this year? Why was now the right time to learn these skills?

In 2021, I took the full-time Codesmith Immersive. It had an amazing hiring program, and I really liked the job I got out of it. I've stayed connected to Codesmith since then, and I saw the opportunity in AI. I'm also getting a master's in AI, and I want to strengthen my technical understanding of the industry and how to use these technologies.

Why did you choose Codesmith’s Technical Leadership Program versus a self-paced, online AI course?

I like the Codesmith teaching methodology, which has many experienced people working in it. I also wanted to work with other engineers and gain knowledge from them rather than do a self-taught course alone. There's more you can get through knowledge-sharing with other engineers than just doing a self-taught journey.

When you were enrolling at Codesmith, what were your career goals? 

I'm a full stack developer who’s been pivoting into AI. Before my current role at Microsoft, I did AI with some startups. With the AI leadership course, I learned the business value of AI, how to integrate AI tools, and what it would cost to integrate those tools into a business model. Before the AI-ML course, I did Codesmith’s Node.js Under the Hood course, which helped my understanding of Node.js, such as how streams and file systems work with Javascript and React.

What was the time commitment like for the technical leadership course? 

The time commitment wavered between 10-15 hours a week. For anyone wanting to take the course, make sure you have the free time to do it! This course is not just about attending meetings. You will also be doing work outside of class time. In addition to the live course hours, you will need 5 to 6 hours a week for projects. For the final project, which we built in a week, we were putting in about 2-3 hours a day. 

Overall, what did you learn in the AI-ML Technical Leadership Program?

It gave me a better understanding of AI technologies and how models work, and I learned which technologies are most dominant in the field right now and how to use them to build applications. Codesmith has a really good program for full stack engineering, and this program helps support it by teaching you how to integrate AI tools into full stack applications. 

What is the difference between learning AI in Codesmith’s Technical Leadership Program versus a master’s program?

In my master’s program, there is a focus on low-level understanding where we read research papers, whereas the Codesmith program is more industry-focused. At Codesmith, you’re figuring out the AI tool to solve business problems and how to integrate it. There are several projects throughout the course that teach you how to build applications with embeddings, LLMs, and vector databases. In the master's program, I’m working more on model-building, train-test-validate data sets, and model-training. 

What were the instructors like for the AI-ML Program?

One of the best parts about Codesmith is the enthusiasm of the instructors. The program has very experienced instructors like Cyrus, who worked with Tesla. I like the overall Codesmith methodology. You can tell the teachers work hard on their lectures. Codesmith is also a camera-on kind of organization. I've taken courses at other places where everyone has their camera off, and it's very uninspiring. Codesmith has a lot of enthusiasm and energy to keep the momentum going.

Tell us about the final project that you built in the AI-ML program!

We created a prediction application that can predict world events. Basically, we got web script data and third-party API data going into a RAG system to LLMs where the user or consumer can ask questions about the likeliness of world events, sports, news, or politics happening.

To get information for this, we scraped data from websites, such as sports websites and news sources. We put that data into a vector database to store all the information as embeddings, and that was our retrieval augmented generation system, or “RAG” for short. In addition to that script web data stored in a vector database, we also had several third-party APIs, such as Kalshi and Polymarket. Those are markets where you can bet on different world events and sports. We added information from those sources as well. 

We took all of this information to a multi-agent system. We had three different LLMs that would evaluate and talk to each other in the Expert Council AI. After they had discussed the issue, they would come up with a prediction, such as predicting the Eagles will win the Super Bowl. 

What were the projects like in the AI-ML Tech Leadership Program?

There are several projects before the final project where Codesmith gives you a repository and you have to integrate embeddings to an LLM or do prompt engineering so it can write proper SQL queries by itself. 

The final project was the team project. In the last week of the program, you're paired up with a team of about 3 to 5 people, and you get to decide what application you want to make.  I was heavily involved with the Pinecone database, whereas another person on my team was involved with the data sources and the data ingestion. This project required us to divide up the work.

Did the rest of your team for your final project have experience in AI before doing this program? 

It was a mix. One common denominator, though, was everyone had a technical background, whether in JavaScript, React, or Python. The program leans heavily on front end and full stack development as a base knowledge.

What was your biggest challenge in putting together your final project?

At this point in my career, I've done several of these kinds of AI projects, so it doesn't have too many new problems that I might encounter. Often the biggest challenge is organizing the work by getting a Kanban or agile process going, assigning work to team members, and sticking to a timeline, even though there's no product manager. For me, the hardest problem with this project was that I wished I had more than a week to work on it! You could keep building up the system, adding different features, etc. if you had the time.

Did Codesmith have a demo day for these final projects in the AI-ML program?

It’s important that Codesmith has a demo day because it builds camaraderie and community. It helps you network with the other students in the program. After the program, having community is something I look for in programs and courses.

What kinds of projects are you working on now as a software engineer at Microsoft?

I am on the PC gaming team on Xbox, which is in a growth stage. In PC gaming, we have over 11 million daily users in over 100+ countries, and Microsoft’s gaming section revenue is over $15 billion dollars. Microsoft released news last month that we will be putting more effort into handheld devices, so a lot of my work consists of supporting that in addition to PC and user engagement.

As far as tech stack, we are leaning heavily into React Native for Windows. Over the last few years, React Native has been building momentum and the Xbox organization at Microsoft is continuing to support the performance benefits. 

Are you using what you learned in the Codesmith Technical Leadership Program now on the job at Microsoft?

Right now, I’m preparing to pivot into more of an AI role, and I’m working with my manager on this. The Codesmith AI-ML program and the AI Master’s program, which I’m finishing in May, are part of a larger career goal to pivot. This year, I will be switching to AI within either the XBox industry or another area of Microsoft that's doing cutting-edge research in AI. 

Was Microsoft interested in your Codesmith experience?

Microsoft is interested in what they're building, but my manager is interested in my Codesmith experience because it makes me a better employee for different projects. If we have something involving AI, that work is going to default to me because I've prepared for it. It's about learning and growth so when the opportunity arises, you're in the right place.

Are you also looking to get any industry certifications to continue helping you on this path?

In addition to getting my Master’s degree, I have a certification from Nvidia and CalTech. The certifications help my career, and they help prepare me for opportunities but I do very much enjoy the personal learning. 

I'm always trying to plan 3-5 years ahead. For example, my first job was originally with Microsoft, and I worked there for a year and a half. I left to work with startups that were integrating AI. I started the Master’s program, and then the Codesmith AI-ML Tech Leadership program. While I was in the Codesmith program, I got rehired with Microsoft!

What are the benefits of doing the AI-ML Technical Leadership Program after completing Codesmith's Software Engineering Immersive Bootcamp? 

Codesmith talks a lot about being a T-shaped developer, which means having solid, broad experience and then having a select depth of experience in one area, whether that’s React, SQL, database management, etc. Incorporating AI, which is such a booming part of the industry right now, helps expand a bit of depth in those broad areas that companies seek. Companies like to see that someone can use data to gain business value insights that can benefit stakeholders. This program definitely makes people more valuable because they can use data for insights rather than just building an application.

What is your advice on how to make the most of this technical leadership program? Is there anything you wish you knew before day one of the program?

Something I've had to learn in the industry is that networking is important to help line up opportunities and gain more knowledge. For instance, when I went to Nvidia GTC last year, I wasn't trying to network, but I gravitated toward people I like, which is basically what networking is. Now, I'm meeting some of them in Yosemite for hiking before this year’s GTC. Don't forget to network as you build up your strengths. Find events whether or not you're in the job search. If you are in the job search, try to do in-person events. I work remotely, but I was going for a hybrid or in-person role because I think there's a lot of value being in-person.

Find out more and read Codesmith reviews on Course Report. This article was produced by the Course Report team in partnership with Codesmith.