Abdus Saboor Dar

Data Science Expert

Evaluated AI model outputs on data science and machine learning tasks, checking statistical reasoning, code correctness, and analytical conclusions. Wrote detailed quality assessments to train and improve models.

Data ScienceMachine LearningStatistical ReasoningCode ReviewModel Evaluation

Data Engineer

IPinfo

Core engineer on IPinfo's Places product, owning the pipeline end to end from ingestion through ETL, orchestration, deployment, and validation. Designed custom ETL over planet-scale observation data and built geospatial matching against global place datasets.

ETLBigQueryAirflowKubernetesGCPGeospatial analysisData ValidationCI CDPythonSQL

Data Scientist II

Afiniti

Built and optimized agent-customer pairing models for Telefónica Spain, scoring live contact-center interactions to select the highest-value match in real time. Measured model impact through alternating hourly benchmarks and statistical inference.

Machine LearningStatistical InferenceA B TestingFeature EngineeringPythonSQL

Dean's Fellow

Delivered lectures and mentored undergraduate students in foundational algorithms and problem-solving. Contributed to research in mathematical and computational methods applied to machine learning and data science.

TeachingMentorshipAlgorithmsMachine LearningData Science

Machine Learning Consultant

Applied text processing and supervised learning to automate SKU matching, reaching 87% accuracy and improving inventory precision by 24%. Engineered a demand forecasting module with XGBoost and built an employee scheduling module using linear programming.

Machine LearningText ProcessingXGBoostLinear programmingPythonSQL

Assistant Manager - Enterprise

Coordinated research projects with external partners, leading teams of 10+ and building SQL-based data models, a data warehouse, and automated reporting. Managed 500+ enterprise client accounts and acquired 100+ new clients while meeting revenue targets.

SQLData ModelingData WarehouseReportingClient ManagementB2B SalesTeam Leadership

Core Team Member

Built Wicktron, an IoT-based no-ball detection system for cricket, taking it from prototype to 85-90% field accuracy through sensor data pipelines and ML model training. Designed the sensor data collection and labeling workflow with validation procedures.

IoTMachine LearningSensor DataData PipelinesPython