Radical Scale

Radical Scale is a regular feature that explores key strategies for growing early stage technology companies.

This issue of Radical Scale looks at the cost of building a machine learning team in Toronto with Radical’s Director of Talent, Meg Lizza.

Increased demand has also brought competition to Toronto, especially for top tier machine learning talent. Within the AI ecosystem, there are two primary factors impacting the AI talent marketplace as US-based AI labs expand their presence in the city and an influx of capital fuels a growing AI startup community. As a result, compensation benchmarks for the best AI candidates continue to climb.

For AI startups, understanding what it costs to recruit the best talent from day one will maximize your company’s ROI and save time and money in the long run. It will also help with retention and create a snowball effect for continuously attracting good people. One or two strong machine learning hires can attract others to join.

I have spent the past three years recruiting and placing top machine learning talent across the Toronto ecosystem. Based on this experience, I have observed compensation bands for tech talent. These are based on actual placements, not compensation surveys from large recruitment agencies or compensation survey companies which do not accurately incorporate the various nuances of machine learning recruiting and technical experience.

The following base salaries are in Canadian dollars and do not include bonus structures or equity benchmarks which differ based on company stage, industry and size.

If you are planning on building out an initial ML team from scratch, this should be your first hire. You want someone who is still close to the code, hands on, but has the EQ to lead and scale a small, dynamic team and present the product/model to internal and external stakeholders. This person will most likely have a Master of Science or PhD, could have a publication record (though not necessary) and has experience deploying AI products into the real world. Getting this hire right is critical and if you do not, it will cost you tens of thousands to fix. My advice: you get what you pay for.

The next hire you’ll need to make is an applied machine learning scientist or research engineer. This hire has a large compensation band. And what you ultimately spend depends on what your company is building as the cost of this hire is closely correlated to their industry experience, academic background, publication record, and ability to write production level code.

For example, a research engineer coming out of the Master of Science in Applied Computing program at the University of Toronto with both research and engineering abilities will typically start at $125-140k base. If your company requires finding someone with top tier research skill sets and implementation skills, then you’ll probably require a candidate with a graduate degree. If you are looking to hire a PhD with a proven first-authored publication track record at top tier conferences such as NeurIPS (Neural Information Processing Systems) or ICML (International Conference on Machine Learning), plus 2-3 years of industry experience and an ability to implement deep learning models, then you are going to be looking at anywhere from $155-200k+ base. It is not required that everyone on your team have a PhD, however, it depends on what your AI roadmap involves. It is not uncommon to see early machine learning teams bring on 1-2 applied scientists with a PhD background.

An ML engineer is someone who can deploy machine learning/deep learning models into production and improve the efficiency, stability and scalability of these deployed models. Candidates with 1-3 years of industry experience post Bachelor of Science or Master of Science start at $110k, while more senior candidates with 5-6+ years of industry experience in infrastructure and machine learning will be closer to 170k base.

The competition for top talent is fierce. While these compensation bands may stretch budgets, be mindful of the ultimate contribution these hires deliver to your company’s AI roadmap and IP. If you are excited about a candidate, consider presenting a strong offer. While it’s important to have some wiggle room in negotiations, low balling risks inviting a bidding war with other companies. If you want to attract the best people in the ecosystem or recruit out of other tech hubs, like Silicon Valley, you will have to pay accordingly.