Dite D.

Key Skills

Software

Don't disclose

Other

Top Subject Matter

Autonomous driving computer vision (perception)

Clinical NLP for medical terminology extraction and NER

Retail product recognition (computer vision) and dataset quality

Top Data Types

Image

Top Task Types

Segmentation

Bounding Box

Freelancer Overview

Senior Data Labeling Engineer | AutonomousDrive AI. Brings 9+ years of professional experience across complex professional workflows, research, and quality-focused execution. Core strengths include Internal, Proprietary Tooling, and Don't disclose. Education includes Master of Science, University of Washington (2018) and Bachelor of Science, University of California, Berkeley (2016). AI-training focus includes data types such as Image, Medical, and DICOM and labeling workflows including Segmentation, Entity (NER), and Bounding Box.

Labeling Experience

Senior Data Labeling Engineer | AutonomousDrive AI

ImageSegmentation

Led a team of annotators to produce 2M+ high-quality labeled images for autonomous driving perception models while maintaining 98.5% inter-annotator agreement. Designed and implemented an active learning pipeline to reduce labeling costs by 35% while preserving model performance. Built custom annotation guidelines and QA workflows for edge cases such as occlusion, adverse weather, and nighttime driving. • Integrated SAM (Segment Anything) into the annotation workflow to improve segmentation efficiency by 60%. • Collaborated with ML engineers to refine labeling schemas using model error analysis. • Improved edge-case model accuracy by 42% through targeted labeling and quality checks. • Reduced pedestrian false positives by 30% via schema iteration and annotation optimization.

2022 - Present

Data Annotation Lead - NLP | MediText Analytics

Don't disclose

Managed annotation projects for clinical NLP models processing 500K+ medical records using HIPAA-compliant workflows. Developed entity labeling schemas for medical terminology extraction (ICD-10 and SNOMED CT) and achieved 94% F1-score on downstream NER tasks. Implemented consensus-based labeling and adjudication to reduce annotation errors by 50%. • Trained and onboarded 20+ domain expert annotators (physicians and nurses) on specialized medical guidelines. • Created automated quality dashboards in real time using Python and Grafana. • Standardized annotation processes to support consistent clinical entity labeling across projects. • Coordinated labeling and adjudication to improve label reliability for production model training.

2020 - 2022

AI Data Specialist | VisionTech Solutions

OtherImageBounding Box

Annotated and curated 1M+ images for retail product recognition systems supporting deployment across 500+ stores. Built Python scripts for automated data validation to detect inconsistencies in bounding boxes and class labels. Conducted data bias audits and rebalancing strategies to improve model fairness metrics across demographic groups by 28%. • Defined labeling requirements for new product categories in collaboration with product teams. • Reduced time-to-market by 20% by aligning label specifications with rollout needs. • Ensured dataset quality through validation checks and consistency enforcement. • Supported iteration on labeling outputs to improve downstream model behavior.

2018 - 2020

Education

Work History