Pushkar Garg

Data Annotator and AI Trainer with expertise in high-volume multi-modal annotation (text, image, audio) achieving 98%+ QA pass rates across 50,000+ labeled samples. Specialized in complex annotation workflows, quality assurance, and cross-functional collaboration with ML engineering teams to deliver production-ready training datasets.

AI Data Trainer & Quality Assurance Specialist at CogniLab AI (2024-01 – Present)

• Annotated 50,000+ multi-modal samples (text, image, video) across NLP, computer vision, and audio ML projects, maintaining 98.5% QA pass rate with zero critical errors over 12-month period
• Achieved top 5% performer status among 2,000+ annotators by delivering 300–500 high-quality labels/day while reducing rework by 35% through meticulous guideline adherence
• Specialized in complex annotation tasks: 3D bounding boxes for autonomous vehicle datasets, NER for medical/legal documents, and multilingual sentiment analysis
• Conducted quality audits on 10,000+ peer annotations, identifying systematic errors and providing feedback that improved team-wide accuracy from 91% to 96% within 2 months
• Collaborated with ML engineers to resolve 200+ ambiguous scenarios, resulting in updated guidelines that reduced inter-annotator disagreement by 40%
• Documented 50+ edge-case patterns with visual examples, decreasing new annotator onboarding time from 2 weeks to 5 days

Software Engineer Intern at CodeAsia (EdTech Nonprofit) (2024-11 – 2025-10)

• Developed Python/Django automation tool eliminating 10 hours/week of manual data entry for 5-person program coordination team
• Built real-time reporting dashboard reducing project status meeting time by 40% and improving stakeholder transparency

Bachelor of Technology in Robotics and Artificial Intelligence – Thapar Institute of Engineering and Technology (2020-07 – 2024-06)