Machine learning and AI have taken the tech world by storm, becoming integral to numerous industries. Working with these technologies daily brings unique challenges and opportunities, as shared by a machine learning engineer who has navigated a career path from physics to data science to machine learning engineering.
Charting a Unique Path
The journey started early for this engineer. At age 12, driven by enthusiasm from The Big Bang Theory, they decided to pursue physics at university. This ambition was a mix of family pride and personal interest, setting the stage for a future in science.
Middle school years were marked by potential yet untapped. The engineer had the aptitude but lacked the discipline. Part of this was due to a youthful arrogance, believing success came effortlessly. As results day approached for university applications to prestigious institutions like Oxford and Imperial College, reality struck. The engineer fell short, facing the harsh reality of not meeting offers and entering clearing.
Clearing in the UK is a situation where universities offer places to students who missed their original offers but have space on certain courses. This second chance led the engineer to the University of Surrey, where they earned a first-class master’s degree in physics.
Despite initial plans for academia, the engineer found that a career in research did not align with their vision. The slow pace and limited opportunities in research pushed them towards data science. One pivotal moment was the release of DeepMind’s AlphaGo—The Movie documentary, sparking interest in AI and machine learning.
Deeply fascinated, the engineer delved into data science, dedicating countless hours to coding, taking courses, and working on projects. This fervent pursuit led to an acceptance into a data science graduate scheme in 2021, formally starting their journey in the tech world.
For more insights, listen to a podcast interview.
Data Science: First Steps
The initial role was at an insurance company, focusing on building various supervised learning models. Tools like CatBoost, XGBoost, and Generalized Linear Models (GLMs) were central to predicting fraud, risk prices, number of claims, and average claim costs. Developing models in a statistical and analytical environment honed the engineer’s skills in regression and classification.
However, the opportunity to advance these models in production was limited due to organizational constraints. The engineer recognized the need to broaden their skill set, particularly in software engineering practices—such as version control, unit testing, and continuous integration and deployment (CI/CD) pipelines—to truly impact the business.
Working with software engineers, the engineer began to bridge this gap. After approximately two and a half years, they became an expert in forecasting and combinatorial optimization. Their expertise helped in both modeling and deployment, earning them a promotion to a more senior data science role.
Transition to Machine Learning Engineering
The engineer’s perspective on data science evolved. Recognizing the value of deploying models in real-time, they aligned with the philosophy that simpler, deployable solutions often deliver more tangible business benefits. Inspired by Pau Labarta Bajo’s quote, “ML models inside Jupyter notebooks have a business value of $0,” the engineer aimed to build and deploy solutions that brought financial value to the company.
This realization guided the engineer towards the engineering side of data science, where they could build and deploy models. With support from their line manager and self-study, the engineer became skilled in writing production code, system design, and deployment processes. Incidentally, the company offered them a machine learning engineer position, aligning perfectly with their skills.
To further enhance their capabilities, the engineer took NeetCode’s data structures and algorithms course and started blogging about software engineering concepts. These efforts not only prepared them for the new role but also showcased their growth and dedication.
The transition came naturally through a combination of experience, learning, and persistence. The engineer’s journey demonstrates the importance of up-skilling and adapting to different roles within the tech industry.
Advice for Aspiring Engineers
Based on personal experience, the engineer offers several tips for those looking to transition into machine learning engineering:
- Gain Experience: Building a robust foundation in data science, machine learning, and software engineering is crucial. A few years of experience in either field, combined with self-study in the others, provides the necessary skills.
- Master Production Code: Learning to write high-quality, tested production code is essential for MLEs. Focus on typing, linting, unit tests, formatting, mocking, and CI/CD. Collaborate with software engineers to learn these practices.
- Understand Cloud Systems: Most modern systems use cloud-based architectures. Familiarize yourself with deployment tools that enable models to go live efficiently.
- Command Line Proficiency: Proficiency in the command line is essential for deploying and managing production code. Use resources to build this skill.
- Study Data Structures & Algorithms: Understanding fundamental computer science algorithms is valuable for interviews. Dedicate time to learn these concepts; it is a manageable yet crucial task.
- Master Git & GitHub: Proficiency in Git and GitHub is essential for version control, code reviews, and collaboration. Focus on these skills to succeed in tech roles.
- Specialize: Many MLE roles require specialization in specific areas. Leverage your previous experience to highlight your expertise in fields like time series forecasting, optimization, or general ML.
The theme of these tips is up-skilling your software engineering abilities. This aligns with transitioning from data science to machine learning engineering, where many roles require a strong foundation in both areas.
Securing a machine learning engineer role is indeed challenging. It requires proficiency in a wide range of skills, but with dedication and the right approach, it is achievable.
Summary and Final Thoughts
The engineer’s career path offers valuable insights into transitioning roles within the tech industry. From physics to data science to machine learning engineering, their journey is a testament to the importance of adaptability, hard work, and continuous learning.
For those interested in machine learning and AI, the engineer’s story provides a roadmap. Start with a strong foundation, leverage opportunities to gain experience, and continuously improve your skills. The world of tech is dynamic, and those willing to evolve with it will thrive.
The engineer invites readers to subscribe to their free newsletter, Dishing the Data. Here, they share weekly tips and advice as a practicing data scientist. Subscribers also receive a free data science resume and a short PDF version of their AI roadmap.
Connect With Me
Join the Journey
Whether you’re a seasoned professional or just starting, there’s always something new to learn in the world of tech. Stay curious, keep working on your skills, and share your journey with the community. Your experiences can inspire others, just as the engineer’s story has inspired many.
Share your thoughts, experiences, and questions in the comments below. Connect with the engineer on social media platforms, and don’t forget to subscribe to Dishing the Data for more insights and updates.