Featured article by Veronika Biliavska, Independent Technology Author
AI and machine learning is one of the most valuable technologies for all kinds of industries. Machine learning can be applied to nearly any kind of industry, though currently, data collection is at the top of the mountain.
In this article, we’re going to highlight 3 ways that individuals with experience in programming and machine learning can put their skills towards financial gain.
Create a startup and build a client base
To create a startup in the machine learning industry, you should be prepared with the following:
– Large dataset
– Domain knowledge
– AI skills
– Capital
It’s a competitive market, but there’s plenty of room for new players in the scene. Let’s examine some statistics from CrunchBase:
– 83% of machine learning startups Crunchbase tracks have had just three funding rounds or less with seed, angel and early-stage rounds being the most common.
– Artificial Intelligence-related companies raised $9.3B in 2018, a 72% increase over 2017, according to PwC/CB Insights MoneyTree Report, Q4 2018.
– Artificial intelligence deals increased in Q1, 2019 to 116 deals, up from 104 deals in Q4, 2018 according to the latest PwC/CB Insights MoneyTree Report Q1 2019.
– AI-based marketing patents are the fasting growing global category, reaching a Compound Annual Growth Rate (CAGR) of 29.3% between 2010 and 2018, according to EconSight.
A great example is Affirm, which offers consumer loans at the point of sale. It raised over $700m USD. Another example Algorithmia, which lets developers employ and manage ML models in the cloud, and is more efficient than AWS. A third great example is Peak, which provides data analytics as a service to enterprise customers.
It all boils down to the flywheel effect. More data gets better AI services/products, which means more products/customers, which feeds the cycle.
To test the waters, create a list of things you care about. It could be climate change, poverty, data centralization, political awareness – anything, really. Next, divide these problems into whether they can be targeted for enterprise or consumer markets. Consumer markets would be things like peer to peer AI computing (for data centralization), and educational chatbots (for poverty / climate change).
For enterprise markets, the focus is on high-value use cases, so for example things like smart C02 emission optimization tools for factories, or cancer detection tools for hospitals.
Next you can create a mock website with a price point and a signup form. If people actually begin signing up, it’s a good sign, and you can begin collecting a client list / relevant data set, and get to work training a model on it. Once you have your model, you take on clients and slowly build your reputation.
Outsourcing your machine learning model
Many companies are doing this, so while it may be a competitive field, there’s definitely room for growth in the industry. Outsourcing your machine learning model is exactly what it sounds like, though most companies are interested in ML that revolves around data collection and insights.
Many startups and SMBs do not have the resources or technical know-how to build an in-house machine learning model, but still have a great need for customer data collection that gives them insight into their consumers’ habits. It can be highly lucrative to offer your machine learning model, tailored to fit a companies specific needs.
Of course you want to protect your intellectual property, and when we’re talking about neural networks and machine learning, companies will be dealing in highly valuable data. A formal contract that outlines a strict agreement between you and the company using your ML model is highly advisory, fortunately you can download free contract templates and have a lawyer review it for any possible loopholes.
Automated Trading Bots
You can create a high frequency trading bot that focuses on cryptocurrencies. Cryptocurrency is a great starting point for a trading bot, because you won’t really be competing against much larger trading firms who don’t deal much in crypto speculation. The technology is still relatively new, considering.
People have made profitable cryptocurrency trading bots, but no one is giving away source code. It’s too valuable to become open source. That doesn’t mean there aren’t resource materials out there to get you started.
The best thing to do is go on GitHub and search for relevant crypto trading bot repos, until you find well documented examples and learn from them. This article also gives some great tips and guidelines on building a bot using deep reinforcement learning.
About the Author
Veronika Biliavska is an independent copywriter. She is passionate about rocket science and ancient Greek literature.
