Pushkar Garg

At CogniLab AI, I’ve annotated 50,000+ text, image, video, and audio samples for NLP, computer vision, and audio ML projects while maintaining a 98.5% QA pass rate with zero critical errors.

I specialize in 3D bounding boxes for autonomous vehicle datasets, medical and legal named entity recognition, and multilingual sentiment analysis. I’ve also audited 10,000+ peer annotations, helping improve team-wide accuracy from 91% to 96%.

I work closely with ML engineers to resolve ambiguous annotation scenarios, improve guidelines, and document edge cases. My documentation reduced inter-annotator disagreement by 40% and shortened new annotator onboarding from two weeks to five days.

Alongside AI data work, I built a Python/Django automation tool and real-time reporting dashboard for CodeAsia, eliminating 10 hours of weekly manual work and reducing project status meeting time by 40%.