At its core Gizmos is a resource management engine building game thus the goal is to maximize your obtainment and value of each resource - chiefly Energy (and indirectly engine power) in the early game and then points in the late game. So naturally you should be focusing on the Gizmos that help you get more Energy and then more points.

Early Game

You start limited to taking one action per turn (technically two with the starting file trigger). Your goal is to gain extra actions as quickly as possible. Buying upgrades is generally a very good idea at this point also. Buying a converter can be good, but for the most part you should be looking for Build into Pick triggers > Pick into Draw triggers > File into Pick triggers. Once you have one or more of these you should look for other Gizmos that align with the color. If you already have a Build Red into Pick Gizmo, buying a Build Red into VP Gizmo is fantastic. If you already have a Pick Blue into Draw trigger, buying a Blue Converter is a great idea. That being said, dipping into other colors is not a bad idea if that's the only option available to you. It's good to be flexible so that you're not only looking for a single color of Gizmo/Energy.

Your first action should almost always be filing a Gizmo. If there are no Build or Pick Triggers in row I or II, then research Level I to look for one.

The File Strategy

It's worth mentioning that some people believe there to be two strategies in Gizmos: one focused on building a lot and another focused on filing a lot. A well played build strategy will beat a file focused strategy more than 90% of the time because building a lot of Gizmos ends the game so much faster than filing. Every file action you take is one less build that you could have taken. The only chance the file strategy has is to build 4 Level III Gizmos but that rarely comes together. Putting your opponents on a clock is very important.

Mid Game / General Strategy

If you have a good build/pick engine going, the value of buying Upgrades, Build into VP, and Converters goes up.

The most skill-testing part of Gizmos is determining when you should take a risk by Researching versus accepting a potentially suboptimal Gizmo from the row. This is something you simply have to build an intuition for over time, but it's worth mentioning a general principle present in all games that involve mitigating randomness: how risky you are playing should be inversely proportional to how much you are winning. Are you winning? Don't take unnecessary risks - just keep winning. Are you losing? Take some big risks to get back in the game - you'll just lose if you don't. Obviously you're usually somewhere in-between but the principle is an important one.

Keep in mind that Level 3 multi-color Gizmos trigger ALL of your Build cards. Building random VP Gizmos even if they do not align with your usual build colors can be good for this reason. But also remember that each game includes a random 16 Level III cards out of the total pool of 36. So it is entirely possible, albeit unlikely, that none of the multi-color Gizmos made it into the Game. There's no way to know until you've researched the whole deck.

Late Game

Reevaluate where you are at point-wise relative to your opponents. Do you need to research Level III hoping to find a big finisher? Or can you just build that face up Level II and coast to a victory? Most of the time the player that builds the 16th Gizmo is also the winner. Double check who first player is so you know who will get another turn and who does not. Try to be the player that sets the pace of the game.