O modelo retorna resultados bem diferentes para capitalizações diferentes.

import requests

URL = "https://g1ad0q0pm5.execute-api.us-east-1.amazonaws.com/prod/predict"
NAMES = ['darci', 'cameron', 'angel', 'morgan']

for name in NAMES:
    for casing in [name.lower(), name.upper(), name.title()]:
        data = requests.get(URL, params={"name": casing}).json()
        print(f"{casing: <12} {data['gender']: <6} {data['score']}")
    print("---")

darci        female 0.6232553124427795
DARCI        male   0.8070128560066223
Darci        male   0.9722667932510376
---
cameron      male   0.9334769248962402
CAMERON      male   0.9525786638259888
Cameron      female 0.5913130640983582
---
angel        male   0.8060540556907654
ANGEL        male   0.861602783203125
Angel        female 0.9087952375411987
---
morgan       male   0.9123147130012512
MORGAN       male   0.941457986831665
Morgan       male   0.9581500887870789