Open to opportunities

Saniya Shinde

@saniyashinde

Message

Data scientist specializing in NLP, computer vision, and ML-driven pipelines.

United States

Message

What I'm looking for

I seek roles where I can build scalable ML/NLP/computer-vision systems, work cross-functionally in Agile teams, focus on production reliability (MLOps/AIOps), and continue growing technical depth and impact.

I am a data scientist with hands-on experience building NLP, computer vision, and predictive analytics pipelines across manufacturing, public policy, and healthcare domains.

Over the past two years I delivered solutions that improved data accuracy by up to 70%, cut manual review by 30+ hours per week, and deployed cloud-based models for real-time inference.

My work includes developing LLM topic-modeling pipelines, large-scale ETL and OCR workflows, and vision-language systems, using tools like PyTorch, TensorFlow, scikit-learn, Docker, and AWS.

I collaborate cross-functionally in Agile environments, emphasize reproducibility and model reliability, and am eager to expand my expertise in AIOps and scalable ML operations.

Experience

Work history, roles, and key accomplishments

Data Science Fellow

The World Bank Group

Jan 2025 - May 2025 (4 months)

Led development of an LLM-based topic modeling pipeline analyzing 25K+ protest records, improving information extraction accuracy by 20% and enabling analysts to focus on policy-relevant insights.

LLM Topic modeling Python Hugging Face AWS EC2 Prompt Engineering Git

Data Scientist (Researcher)

George Washington University

May 2024 - May 2025 (1 year)

Developed NLP pipelines processing 250K+ unstructured government documents, reducing manual review time by 30 hours/week and building large-scale Python ETL workflows with OCR.

NLP Python OpenCV ETL Data Engineering PDF Parsing

AI Research Engineer

Independent Project

Feb 2025 - Mar 2025 (1 month)

Developed end-to-end VLM system to generate structured medical reports from 7K chest X-rays, fine-tuning BioGPT with LoRA and deploying via Docker on AWS for real-time inference.

BioViLT LoRAs PyTorch Docker AWS EC2 Streamlit

Data Scientist

Bimcon Inc.

Jun 2024 - Aug 2024 (2 months)

Automated cleaning and preprocessing of 315K+ inventory records, improving safety and compliance decision-making by 50% and built models to predict flammability indicators.

Python scikit learn AutoML SeaBorn Matplotlib Data Cleaning

Data Analyst

Yardi Software

Sep 2020 - Jul 2023 (2 years 10 months)

Built automated Power BI and SSRS dashboards and 200+ MySQL procedures to reduce vacancy losses, saving $120K annually and improving reporting efficiency via query tuning.