A manager of Cardiff University's coronavirus testing system quit as he felt it was "not trustworthy".
Neil Evans said he felt the system was rushed and that it could result in false-negative results, meaning potentially infectious students could be missed.
Cardiff University said it strongly rejects Mr Evan's claims, adding that "hundreds" of asymptomatic cases have been identified.
The Cardiff University system has carried out around 25,000 tests since it started, with the institution claiming "a few hundred" asymptomatic positive cases were identified.
Students and staff would provide a saliva sample, via a swab, which then underwent a PCR test.
These tests look for the generic code of the coronavirus. It confirms if an individual currently has the virus, but cannot confirm if they have had it previously.
Any positive cases identified were then confirmed by an NHS test.
Mr Evans was a service manager in August and September when the system was being developed. He resigned on September 23, just two weeks before the programme was launched.
Mr Evans said after several years as a researcher he was approached to become the service manager to set up the service.
At first he said he felt pleased to be part of the project, but felt it "became quickly apparent" the system would not work.
He explained that the results they collected needed to be tested against known samples from an accredited NHS lab, But Mr Evans said they needed more time to validate the setup.
Mr Evans explained that when a new system is set up, known positive and known negative samples will be sent from an accredited lab, usually an NHS source.
These are then used as a comparison to make sure any new tests, in this case those done by the university, match up. This ensures that negative results are actually negative and the same for positive results.
A university spokesperson added that the university estimates their tests agree with the NHS test around 90% of the time.
Referring to the 90% figure, Mr Evans said the number "isn't great" and said the university has not expanded on what the total figure that is a percentage of.
He added: "I don't think the system is trustworthy. I don't think it's fit for purpose. They need to do a lot more work. Which they do say they are working towards accreditation, but to me, they shouldn't be giving any results to anybody until they absolutely 100% know that this negative result really is negative."
Mr Evans described the system as a "catalogue of errors". A university spokesperson said their "validation processes are vigorous and thorough".
He said: "They were basically out of their depth from a clinical diagnostic perspective. I've worked for the NHS for over a decade so I am an expert in this field.
"In the end I thought I'm not going to put my name to this. It would have been me, as the service manager, who was overseeing the results and issue them. The biggest issue for me is the false negative. Without a through validation you have no handle on what your lower limit is.
"You could be calling someone positive in a lower level negative, and that's obviously really bad."
President of the Institute of Biomedical Science, Allan Wilson, also shared some concerns of the system.
On Tuesday he said: "One concern was there was no evidence of how they validated the test when they introduced it, by that I mean an internal validation."
He explained that normally in a lab when a new test is introduced a known set of tests already validated from another lab, so confirmed negative results, for example, would be run to confirm outcomes from that record to compare. Ongoing interim tests would also normally be carried out, he added.
Mr Wilson said that the use of saliva with this type of test is "unusual".
He said: "The fact they are using saliva for the test is unusual. I don't know of any other lab that's using saliva as a sample on a PCR test.
"That's what makes that validation step even more important, the fact they are doing something different from most other labs."
He said: "The fact they are using saliva for the test is unusual. I don't know of any other lab that's using saliva as a sample on a PCR test.
"That's what makes that validation step even more important, the fact they are doing something different from most other labs.
"There is that potential risk [of false negatives], but there is that potential risk with any test. No test is 100%.
"But what you have to do is put as many steps in place to reduce the risk of both false negatives and false positives."
A Cardiff University spokesperson said: "Cardiff University invested, and continues to invest, in an asymptomatic (those without symptoms) screening service for the sake of the health and wellbeing of our staff, students and wider community.
"The overriding aim of our screening service is to give the University and Public Health Wales (PHW) the intelligence needed to help us to take action to prevent the spread of COVID-19 among staff, students, subsequently their families and the wider community.
"The screening service was never intended to be a symptomatic testing service, and all positive results are confirmed by an NHS test.
"Our screening service led to the successful identification of hundreds of asymptomatic (no symptoms) cases – then confirmed by an NHS test - and led to those students self-isolating. Without the service, they would have been carrying the virus and spreading it unknowingly, with severe implications for the student and local Cardiff community."
The university said any member of staff or student who have symptoms should be tested by the NHS, in the usual way. They also release data on testing numbers and cases here.
The Wales infection rate is now 102.3 per 100,000 population for the seven days up to February 7, a decrease from 107.4 on Thursday. It has not been this low since September. You can read more here.
