What Professional Machine Learning Engineers ACTUALLY Do

Summary

This video script offers a candid look at the daily life of a machine learning engineer or data scientist, debunking the glamorous portrayals often seen online. It emphasizes the reality of long hours behind a computer, the collaborative yet solitary nature of the work, and the significant time spent on documentation and dealing with messy, real-world data. The speaker candidly discusses the challenges and the iterative process of refining models amidst noise and unpredictability, while acknowledging the job's rewards despite its imperfections.

Takeaways

• 🖥️ The primary activity of a professional engineer, such as a Machine Learning Engineer or Data Scientist, involves sitting behind a computer for at least 6 hours a day.
• 🔬 The 'Day in the Life' videos often show glamorous aspects but omit the more mundane, solo tasks that dominate the actual work life.
• 📅 A typical day includes a stand-up meeting to discuss progress, issues, and plans with the team, which can vary in length and interaction level.
• 🤔 The role often involves helping others with infrastructure-related problems, indicating a collaborative and supportive team dynamic.
• 🔍 The job can be highly specialized, with engineers focusing on a specific type of machine learning problem or domain for extended periods.
• 🏭 Larger companies may offer a broader range of tasks, but even then, an engineer is likely to become an expert in a particular area and handle similar projects.
• 🛠️ Engineers are responsible for maintaining and working within a specific code base, cloud, programming language, and tools, which might not change frequently with trends.
• 📝 Documentation is a significant part of the job, including writing manuals, emails, and updates on platforms like Confluence, which can be time-consuming.
• 🗑️ Dealing with messy, real-world data is a major part of the job, often involving cleaning and making sense of noisy and incomplete data sets.
• 🔄 Projects may not always have a clear outcome or may not work as expected due to the inherent noise and unpredictability of real-world data.
• 🌐 Despite the challenges, being a Machine Learning Engineer or Data Scientist is still considered one of the best jobs in the world, offering a rewarding career.

Q & A

What is the common misconception about the life of a data scientist or machine learning engineer portrayed in videos?

The common misconception is that their life is filled with exciting activities such as attending conferences, hackathons, visiting fancy coffee places, and going to the gym, which is not the typical daily routine.

What is the reality of a professional engineer's workday according to the script?

The reality is that a professional engineer spends a significant amount of time sitting behind a computer, often for at least 6 hours a day, and their work involves coding, meetings, and dealing with various technical and business-related issues.

What is the typical start of a day for a machine learning engineer?

A typical day usually starts with a stand-up meeting where the team discusses what was done the previous day, what is planned for the current day, and any problems they are facing.

What does the script suggest about the structure of teams in a machine learning engineer's workplace?

The script suggests that team structures can vary, with some having multiple people in the same role and others being cross-functional, where the machine learning engineer might be one of the more technical members compared to data analysts or data scientists.

What are some of the less glamorous tasks that a machine learning engineer might be responsible for?

Some less glamorous tasks include helping with infrastructure issues, dealing with support cases, and engaging in technical discussions about APIs, computing clusters, storage technologies, and data protection.

Why might someone considering a career in machine learning engineering want to ask about different data roles in a company during an interview?

Asking about different data roles can help a candidate understand the specific responsibilities and expectations associated with each role, as well as the level of variation they can expect in their work.

What is one surprising aspect of the job that the script mentions?

One surprising aspect is the amount of documentation work involved, such as writing pages on Confluence, emails, and teams messages, which can take up a significant portion of the workday.

How does the script describe the nature of data that machine learning engineers typically work with?

The script describes the data as mostly noise and garbage, often requiring extensive cleaning and understanding of why certain anomalies exist in the data.

What is the script's final message about the job of a machine learning engineer?

Despite the challenges and potential misconceptions, the script concludes that being a machine learning engineer is still one of the best jobs in the world, offering a lot of satisfaction for those who love the work.

What advice does the script give for someone who might be considering a career in machine learning engineering?

The script advises considering one's personal preferences and lifestyle, as the job can involve a mix of machine learning project work, infrastructure work, support, and meetings, which may fluctuate depending on various factors.

What does the script suggest about the importance of understanding the specific tasks and domain of a machine learning engineer in a company?

The script suggests that understanding the specific tasks and domain is crucial because a machine learning engineer may end up working on the same type of problem or within the same domain for an extended period, which can define a lot of their time and work.