Code › codeit-ai-sprint
Leaving the Codeit Sprint Early for a New Job
Closing my Codeit AI Engineer Sprint chapter early after accepting an AI agent engineering role
Today is the end of Part 1 of the Codeit Sprint, and it is also my last day in the program. I accepted an AI agent engineering role at a financial AI company.
People around me said that going back to work was the obvious choice, but I spent a long time thinking about it. I had treated the AI engineering bootcamp as a chance to push my skills as far as I could during my time away from work. It reminded me of the post I wrote during the pre-course about why I joined an AI bootcamp after twelve years as a developer.
The New Job and My Last Day
I worked really, really hard at the bootcamp. I joined every opportunity that was available, and I wrote about what I learned each day so I could digest it and review it again later.
The energetic lead of Team 2 and I even organized a study group around the remastered edition of Deep Learning from Scratch 1. Unfortunately, I had to leave before our first meeting. I felt sorry about that, though another teammate was able to take my place.
Outside Codeit, I was also taking part in a separate algorithm study group. Doing both left me completely exhausted, and I sometimes wondered whether I had taken on too much. Still, algorithm skills fade without regular practice, and I did not want to miss the next opportunity because I had stopped preparing. I am still keeping up with that study group now.
That connects to something I have learned from interviewing with several global big tech companies over the years. I once reached the final round at one company’s Seattle headquarters and lost the opportunity at the last step. Repeatedly going through that process taught me that preparation has to continue before the next opportunity appears. That was one reason I considered the bootcamp such a valuable chance to strengthen my AI skills.
There is another side to the decision. Learning by working through real problems in production tends to stay with me, and it also becomes part of my professional experience. I considered the financial AI domain, the agent engineering role, the people I would work with, and the terms of the offer together before deciding to join the company. I also decided that returning to work would not mean stopping the areas I still wanted to develop.
The Time I Spent in the Codeit Sprint
I studied machine learning and the early part of deep learning. From the pre-course until today, I turned almost everything I studied into a blog post. I certainly have not absorbed all of it yet, but writing one post a day to organize the material is one of the things I will remember most.
The lead instructor had a lot to cover in a limited time, but the classes connected code, data analysis, practical experience, and lessons from production work. I am grateful for that.
Every member of Team 4 also had a distinct personality, which made the team fun. I tried to do a decent job as the team lead, though I am not sure how well I actually did. Haha.
And our mentor was genuinely great. I learned a lot from those sessions, and the mentor listened to many of my concerns about working in the field. The mentoring time felt particularly warm.
The daily log challenge also played a large part in keeping me writing every day. I had participated since the pre-course, and the small sense of competition made it fun. As of today, I am still at the top of the ranking for the number of posts written. I expect the rest of cohort 14 will keep it going from here. It gave me a reason to study, organize what I learned, and maintain my blog at the same time. For me, it was a three-for-one.
What Remains After a Short Month
I only completed one month of a seven-month program, but that month included mission assignments, achievement assessments, and a lot of Codeit lessons. I learned the mathematical ideas behind tools I had previously used without fully understanding what happened inside them.
The machine learning topics included linear and logistic regression, decision trees and boosting, evaluation metrics, and support vector machines. In deep learning, I reached multilayer perceptrons and convolutional neural networks. I also worked through exploratory data analysis and trained models directly in Google Colab.
I promised myself that I would finish my responsibilities as team lead and complete the remaining Part 1 lessons before leaving. I finished those today.
Continuing the Learning in My Next Role
I return to work next week. I am nervous and excited about how I will contribute in the new role. Once I start, learning what the company needs will come first. With the time that remains, I plan to continue the AI topics I could not finish on my own or organize another study group. I also want to keep studying AI inference and serving, which I discussed with my mentor.
Reading this back, it sounds like I am talking about studying, studying, and more studying. I will take that as a sign that I still have plenty of room to grow.