Data Scientist - ML Data Quality Expert

Ayushi P

Baltimore, MD +1-443-***-**** ***********@*****.*** LinkedIn GitHub

SUMMARY

Data Scientist with 3+ years of experience across data science and data engineering. At Amazon, reviewed 1,000+ Alexa+ responses and helped maintain 96%+ accuracy by checking data quality and handling unclear labeling cases. Earlier at Sigmoid, worked with Python, SQL, FastAPI, Azure, ETL workflows, CRM data, Google Drive integrations, and reporting automation. My background gives me experience with both the analysis side of data and the systems behind it. I’m now looking for a Data Scientist role where I can apply that mix to real business problems. PROFESSIONAL SKILLS

Programming & Data: Python, SQL, SQLite, Pandas, NumPy, Data Cleaning, Data Validation, Data Modeling Machine Learning: scikit-learn, Classification, Predictive Modeling, Feature Engineering, Model Evaluation, Stratified Validation, Probability-Based Scoring

Data Engineering: ETL, Data Extraction, Data Transformation, Data Integration, Reporting Automation, Data Quality, Analytics-Ready Datasets

Cloud & Backend: Microsoft Azure, FastAPI, REST APIs, Azure AI Search, API Integration, Caching, LLM Provisioning Analytics & Visualization: Tableau, KPI Analysis, Customer Segmentation, Attrition Analysis, Interactive Dashboards AI Data Operations: Data Annotation, ML Data Quality, Model Response Evaluation, Labeling Quality Control, Alexa+ Data Processing Applications & Automation: Streamlit, Google Drive API, CRM Integration, Workflow Automation, File Processing DevOps & Tools: Azure DevOps, Mermaid, GitHub

Responsible AI: Model Explainability, Data Privacy, Responsible Use of Model Results EXPERIENCE

Data Scientist Intern Amazon USA May 2025 – May 2026

• Applied data annotation and quality processing skills for ML models for 1000+ Alexa+ responses ensuring the accuracy for the site was always 96+%.

• Engineered a Mermaid AI flowchart to mitigate data labeling errors to maintain site-wide accuracy of 96%+, translating complex evaluation requirements into a clearer visual framework for coworkers and professionals when encountering ambiguous data labeling tasks.

• I also had the opportunity to speak with and explore how my role as ML Data Associate gets to the customer speaking with professionals such as a Quality Auditor, Quality Specialist, Language Engineer, Expert Consultant, and Technical Account Manager and see how the life cycle of data goes at Amazon for Alexa+ products, the importance of data, and what constitutes as good or bad data. Data Engineer Sigmoid India Mar 2022 – Jul 2024

• Built and maintained backend APIs using FastAPI and Azure for LLM provisioning, secure access, and usage tracking at the user and app levels.

• Implemented caching with Azure AI Search and automated workflows to reduce model costs, support file uploads, and streamline deployments via Azure DevOps.

• worked on end-to-end automation of data extraction and reporting workflows, integrating Google Drive API, CRM data, and WPSR reports to streamline operations.

• Reduced manual effort by 95% and improved the project performance visibility for senior leadership and stakeholders. PROJECTS

Banking Customer Analytics Python, SQL, SQLite, Tableau, ETL, Data Modeling GitHub: https://github.com/ayushipatel4/Bank-customer-analytics-/tree/main

• Built an end-to-end Python and SQLite ETL workflow to validate, transform, and analyze banking data, using SQL and calculated KPIs to study customer attrition, account dormancy, cross-sell opportunities, and balance exposure.

• Created analytics-ready datasets and an interactive Tableau dashboard with customer segments and branch-level metrics, making it easier to drill into retention risks and revenue opportunities. Medical Triage and Disease Prediction Python, scikit-learn, Streamlit, SQLite, Machine Learning GitHub: https://github.com/ayushipatel4/Medical-Triage

• Built a machine learning workflow for medical triage and disease prediction using symptom and patient data, including feature preprocessing, stratified validation, and comparison of classification approaches.

• Developed a modular Streamlit application with saved models, cached inference, probability-based triage scoring, and SQLite patient- record storage, while accounting for explainability, privacy, and responsible use of model results. EDUCATION

M.S. Data Science University of Maryland, Baltimore County, USA Aug 2024 – May 2026 B.Tech Computer Science and Engineering Amity University, India Aug 2020 – Jun 2024 CERTIFICATIONS

AWS Certified Data Engineer Associate (2025)

Contact this candidate