Q1 2026: A Solid Start to the Year
Planning, DSA, Transformers-based classifiers and a sustainable routine: a review of the first quarter of 2026
2025 In Review
Starting in mid-2025, I began approaching my career development with a more intentional, Inspectable Shipped Hours (ISH) mindset. The first step was launching this website, followed by a reflection on where my professional capital stands, which gaps are most important, and which areas are likely to matter most over the long term.
In the second half of 2025, this reflection prompted me to work across several areas:
- data structures and algorithms on LeetCode
- a cleaning robot REST API (~20 ISH)
- an adversarial prompt classification project with transformers, together with a small Edge AI project (~16 ISH)
- German language study with the Goethe-Institut, resulting in the completion of the A1.1 module (~60 non-ISH hours).
Q1 2026
I started the year by organising my project ideas. I defined 12 projects that can help me build career capital across several areas, including C++, MLIR, SQL, hardware-software co-design, agentic AI, and MLOps.
I also continued working on the adversarial prompt classification with the transformers project. The core had already been implemented in a Jupyter notebook in 2025, so my original plan was to refine the methods to extract additional performance. Instead, I decided to take a different route and refactor the notebook into a production-style codebase in the Transformer-Classifiers repository. Thus, rather than focusing narrowly on model iteration, I spent more time on repository structure, CI, automated documentation with MkDocs, secret handling, and data management using Hugging Face datasets. This is slower, yet more valuable from a software engineering perspective. The project remains ongoing and will continue into Q2.
As a self-taught programmer, I have had relatively little formal exposure to data structures and algorithms. Thus, I continued working through LeetCode’s Data Structures and Algorithms Crash Course, tracking my solutions in my Data Structures and Algorithms repository. Over the quarter, I completed the Arrays and Strings and Hashing sections, made progress on Stacks and Queues, and started Binary Search.
I also wanted to remain engaged with the wider machine learning and data science community in Switzerland. To that end, I applied to attend the Swiss Conference on Data Science (SDS) 2026 in Zürich. Although attending in a personal capacity, I was encouraged and supported by my managers at Synthara to submit a business-talk proposal connected to my work. This was accepted, and I will present a poster titled “Selling the Invisible: A Data-Driven Approach for Enabling In-Memory Computing for Edge AI”.
The final milestone of the quarter was completing my Goethe-Institut German A1 course by finishing the A1.2 module, having spent ~75 (non-ISH) hours.
Career Capital Tracker
| Project | Total ISH | Status | Notes |
|---|---|---|---|
| Adversarial Prompt Classification with Transformers | 24.5 | Below plan | Shifted project scope from refactoring to infrastructure, reduced short-term visible output. |
| Data Structures and Algorithms | 17 | Below plan | Steady progress, but less than originally intended. |
| SDS 2026 submission | 9 | On plan | Proposal prepared and accepted. Poster and write-up to follow in Q2. |
In total, I accumulated 50.5 Inspectable Shipped Hours (ISH) in Q1 2026.
This is below what I had hoped for at the beginning of the quarter. However, it is a good start to the year, particularly given the workload and travel during this period. In addition, I began building a solid routine for my personal development, which is the most important achievement of the quarter.
Books Read
Alongside the projects above, I also read the following books across late 2025 and Q1 2026:
- So Good They Can’t Ignore You by Cal Newport
- Deep Work by Cal Newport
- Digital Minimalism by Cal Newport
- Nexus by Yuval Noah Harari
- How To Be A Stoic by Massimo Pigliucci
- The Obstacle is the Way by Ryan Holiday
I am currently reading:
- Big Trust by Dr Shadé Zahrai and Faysal Sekkouah
- Stoicism For Dummies by Tom Morris and Gregory Bassham
Next Quarter
In Q2 2026, my focus will be on:
- Continuing the Adversarial Prompt Classification with Transformers project, aiming to complete the refactoring work, exploring a small number of targeted optimisations, and publishing a project write-up (~60 ISH)
- Progressing through the Data Structures and Algorithms course, aiming to complete Stacks and Queues, Linked Lists and starting Trees and Graphs (~40 ISH)
- Attending SDS 2026 in Zürich, presenting the poster, and publishing a short post about it. (~10 ISH)
- Completing the A2.1 module of the Goethe-Institut German course and starting A2.2. (~75 non-ISH hours)