I was involved in a bridge project that had clearance issues. Raising the bridge would have involved re-profile-ing a significant patch of Interstate Highway. Lowering the roadway below was limited by groundwater issues (it was only a couple of hundred yards from the Columbia River) and an existing sanitary sewer. So the thing was designed at minimums, to the extent of affecting the chord of the beams used. When constructed it sagged a little more than expected - not so that you'd notice by looking at it - but using up every bit of the minimal safety factor that had been designed in.
These guys blamed the surveyor too!:-D
Although, I worked a RRD project which had lots of room for misinterpretation. The RRD designed a grade separation. And they took on both the 1/2 mile of road design and 2 bridges. One for the train, one for a street. It was a complicated project. The RRD engineers do a fine job on RRD bridges. But they made a lot of design errors in the road portion. Ultimately most of the issues were uncovered and resolved in the construction staking. But there was a lot of redesign in the field. Now, this was a one-off, atypical situation. The RRD should've had a civil eng. When it came to super critical elements, by then, everyone knew that triple checking, and review was the only way to go. That was a long time ago, and CAD was still new. So, it OJT. Fortunately the only major oversight was a bridge deck was built higher than necessary. The bridge deck was diagonal to the street section. And the portion of street involved was a 4% ramp. They carried the elevations across the section perpendicular to centerline, forgetting that the street (and deck) went downhill 2'. The surprise was discovered after it was too late, and it was built per plan. Could've gone the other way.