I'm going to assume the liquid in question has a dielectric constant that is very much greater than the dielectric constant of air. This is the case for many liquids, including all liquids whose major component is water. Many "medical liquids" fall in this category, as people are mostly water too.
I am am going to recommend a capacitive measuring system based on the concept described Scott Edwards' article "Measure Water Level
Without Getting Wet" (Stamp Applications no 27, 1997):

Basically, you put the liquid in a small narrow tube, either as a container or through communicating vessels from a larger reservoir. You apply conductive tape to both sides of the tube. You now have made a capacitor with the liquid as dielectric. Because the dielectric constant of the liquid is much higher than the dielectric constant of air, the capacitance of the system increases with increasing water level. Because of the finite dielectric constant of air, the relation is linear but not strictly speaking proportional.
The good thing about this system is that it's relatively accurate, and the electronics never get anywhere near the liquid.
For your specific case, you are measuring relatively small quantities. If it's acceptable in your application, I would get rid of the reservoir altogether and replace it with a single tube about \$10 cm^2\$ in cross section and one meter long. This should enable you to store 1 liter of fluid while giving you plenty of swing in your capacitor value.
For actually measuring the capacitance, there are many solutions. Here's a few:
- Use a CD4060 as in the basic stamp note.
- Turn the capacitance into a frequency with a schmitt-trigger oscillator.
- Use the cap to make an RC filter, feed it with an AC signal (e.g. a square wave from a microcontroller clock), put the result through a rectifier circuit (using an op-amp based circuit to avoid the diode drop) and measure the resulting average voltage
You can find more circuits to perform the capacitance measurement in the comments of this blog post I made on the subject of capacitive measurement.
Actually, if you just need discrete high/low levels, float switches would work fine, are super reliable, and not expensive. Look for a sump pump type float switch. Basically a mercury switch in a sealed floating ball.
If you need actual level, one of the best systems is a 'bubbler', which uses a controlled air pressure to literally blow bubbles out of a tube, which has the endpoint at the bottom of the liquid stack. The level is measured by monitoring the air pressure and air flow supplied to the bubbler tube. Bubblers are good as they are inherently self cleaning, and give instant indication if there is a severe clog or fault in the system.
Another option for actual level would be ultrasonic. Ultrasonic rangers aren't expensive, and using ceramic transducers, the corrosion problem is eliminated.
Best Answer
Also consider using a capacitive sensor.
If you glue two plates side-by-side on the outside of the tank, they will act as a capacitor with the water (or air) in the tank as the dielectric. Because water has a dielectric constant that is 100x that of air, you can easily and accurately relate the capacity of the plates to the water level in the tank.
See Measuring Water Level Without Getting Wet for more information.