The infrastructre and costs of HFT (hardware, co-location, etc.) are enormously high and as far as I know only institutions can compete on this level. However, if you have an algorithm that has an edge that does not depend on millisecond level execution, you're just talking about automated trading and this is very achievable by an individual.
Many brokers offer APIs to service this market. For example, Interactive Brokers has a commonly used API that has a sizable community of people working with it.
About 6 years ago, I developed an automated trading system with a friend who was a self-employed day trader at the time. It was a very interesting project and quite successful for a while. However, as time passed, markets changed and so did the profitability of our algo. At our peak we were doing 10k profit/month after commissions, but it was short-lived.
There are several communities, such as at IB, or Elite Trader, to discuss the topic of automated trading, but don't expect the helpful atmosphere you get here. It's an extremely competitive environment (even though you're almost never directly competing with the small-time investors who post on those forums) and there is much more misinformation and trolling than there is useful reading.
What you need is an edge of course, but a real edge. The guys who make money for a year or two or five and then lose it all in year six are also common. They thought they had an edge and didnt. The problem of course is, as with arbitrage, the market itself eliminates any edge for it participants. It is possible to be successful, true its not easy, but im not sure what in life thats worthwhile is easy. Guys who say, its impossible, you will lose. Thats ignorance. HFT definately has an edge, but its not as much in their propietary strategies as people think. HFT firms can make on average a tenth of cent profit on every share traded, and be profitable. Of course they trade large volume, and with their broker dealer relationships pay a very very low commission rate, a rate so low its unavailable to the average joe. If they paid what you pay to make a trade, they would go broke. Their other edge, techniques to flash orders in high speed to the exchanges, are in part edges based on their expensive high speed connections, but also in some cases lax regulatory environment. Generally, markets have rules against ..manipulation.. Meaning showing or placing an order they have no intention of executing, but placing it to manipulate the current price and there are a few ways to do this. Exchanges overall have cracked down, but not always to the extent they could. Some trading operations have been caught and fined by the sec for this kind of activity. And then of course their strategies mostly based on statistical methods which presumes the price movements are random and statistics can be used to predict them. With the other advantages, high speed connections, extremely low fees, high volume.. They dont need a crystal ball in their strategy, but something just good enough to turn the 50 percent chance of success to say 50.1 percent.
Recommendations for you, learn by doing and get a low fee broker. Not giving away your profits in commissions is absolutely essential. Interactive Brokers is a common choice for the retail trader and pretty good choice overall. With trading, its tough as you may have to go bankrupt a couple times to learn what works. More than many people are willing to tolerate.
If by HFT, you are referring to quantitative trading then there are several sites like Think or Swim, Ninja Trader, etc that you can use. There is lot of research literature available online for Quantitative Trading ideas.
For the individual you would either need to be very good (best of the best) or do it for a very long time. You would still require a sizeable bankroll.
I'm currently taking https://www.coursera.org/course/compinvesting1, it's quite basic but the next course, 'Computational Investing 2' will go through Machine Learning for investing (scheduled for January I think).
Hope this helps.