I'm a curious Engineering Manager with 19+ years of experience in developing software products and leading people. Having lived in 5 countries, I've learned to collaborate effectively with multi-disciplinary and geo-distributed teams. Having worked across industries such as Retail, Energy, Manufacturing and Gov, enables me to drive business outcomes and engineering excellence in the age of AI.
- Learn multiple ways of solving the same problem and know the pros and cons.
- The term best practices is relative. Best practices constantly evolve over time.
- Never settle in learning in-depth how something works. Embrace unlearning and re-learning.
- Instead of only doing your part, help out in dependent workstreams too to fully understand the E2E. Aim to solve the "whole problem".
- The most dangerous phrase: "we've always done it this way".
- Never be too busy to improve. Don't get stuck in a hamster wheel of inefficiency.
- Avoid unearned complexity by practicing YAGNI.
- Beware of the sunk cost fallacy.
- Out of all the agile practices commonly used, estimating work items and trying to measure project velocity is the least productive.
- Looking at GenAI solely for speed gains misses its deeper value, which lies in decision intelligence, and uncovering quality gaps rather than just doing old tasks faster.
- Put people over processes.
- Assume best intentions.
- Maintain awareness of your own biases.
- Judge people on their contributions, not on how confident they seem.
- Don't let your own desire to get things done quickly, turn into undue pressure on colleagues.
- Success is not a zero-sum game, so be happy and content with others' success.
- Customer is not a monolithic entity, it's a mosaic of individuals with vastly different motivations, habits, and expectations.
- New systems are best designed by a small number of minds, not committees.
- Leaders model calm. When you stay calm under pressure, your team follows suit.
- As a Manager/Lead, if things go well, give your team the credit. If things go sideways, take the blame yourself.
- Code is a liability, not an asset. Each meaningful piece of code has to be version controled, tested, secured, deployed, monitored and documented iteratively throughout it's evolution.
- Code is just one of the many ways of achieving a solution.
- You own the code and are responsible for it, whether it was written by you, copied from Stackoverflow or generated with AI.
- Your team is only as good as the weakest code reviewer.
- Large PRs increase the cognitive load on human code reviewers. Be kind to them by attaching a proof that your PR actually works (e.g. screenshots or short videos).
- Don't get sucked into hyped/fashionable tech. Internalize that CS fundamentals haven't changed much over time.
- Knowledge of specific frameworks, libraries or tools is not that important in the long run.
- Keep the docs as close to the actual source code as possible.
- All code in the critical path should have good tests, regardless if the tests were written first, last, or in the middle.
- Deploy from a main branch to an environment from the very beginning of a project.
- Don't implement in the app code, what can be configured in the infra.
- Automate all the things that are worth automating as early as possible.
- Ensure the system is observable so that you can enquire at any time what is happening.