Looks like you have JavaScript disabled. For the full Twine experience, you will need to re-enable it.

I’m a data science and AI practitioner based in Boulder, passionate about building practical ML solutions for robotics, finance, and enterprise workflows. I enjoy turning messy real-world data into robust models, deploying end-to-end pipelines, and collaborating across teams to translate analytics into measurable impact. In my free time, I explore new LLM capabilities, experiment with multi-GPU training, and contribute to open-source projects. I thrive in remote-first environments and love turning ambitious ideas into tangible products that improve decision-making.…I’m a data science and AI practitioner based in Boulder, passionate about building practical ML solutions for robotics, finance, and enterprise workflows. I enjoy turning messy real-world data into robust models, deploying end-to-end pipelines, and collaborating across teams to translate analytics into measurable impact. In my free time, I explore new LLM capabilities, experiment with multi-GPU training, and contribute to open-source projects. I thrive in remote-first environments and love turning ambitious ideas into tangible products that improve decision-making.

Shreyash Sahare

Data Scientist, AI Engineer, Developer, +2





I’m a data science and AI practitioner based in Boulder, passionate about building practical ML solutions for robotics, finance, and enterprise workflows. I enjoy turning messy real-world data into robust models, deploying end-to-end pipelines, and collaborating across teams to translate analytics into measurable impact. In my free time, I explore new LLM capabilities, experiment with multi-GPU training, and contribute to open-source projects. I thrive in remote-first environments and love turning ambitious ideas into tangible products that improve decision-making.…I’m a data science and AI practitioner based in Boulder, passionate about building practical ML solutions for robotics, finance, and enterprise workflows. I enjoy turning messy real-world data into robust models, deploying end-to-end pipelines, and collaborating across teams to translate analytics into measurable impact. In my free time, I explore new LLM capabilities, experiment with multi-GPU training, and contribute to open-source projects. I thrive in remote-first environments and love turning ambitious ideas into tangible products that improve decision-making.

Available to hire

I’m a data science and AI practitioner based in Boulder, passionate about building practical ML solutions for robotics, finance, and enterprise workflows. I enjoy turning messy real-world data into robust models, deploying end-to-end pipelines, and collaborating across teams to translate analytics into measurable impact.

In my free time, I explore new LLM capabilities, experiment with multi-GPU training, and contribute to open-source projects. I thrive in remote-first environments and love turning ambitious ideas into tangible products that improve decision-making.

Skills

Experience Level

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Work Experience

Robotics Engineering Intern at Void Robotics Marathon

August 1, 2025 - Present

Developed a CV defect detection pipeline using K-Means clustering and OpenCV for analyzing 50+ real-world image sets. Implemented YOLO-based object detection integrated with ROS2, enabling robust real-time perception and object tracking. Integrated SLAM navigation with sensor fusion, demonstrating model deployment in robotics environments.

Data Science Expert (Independent Contractor) at Mercor

September 1, 2025 - October 15, 2025

Executed 20+ ML projects covering hypothesis testing, time series forecasting, and predictive modeling. Crafted prompt engineering workflows for Gemini 2.5 Pro, benchmarking reasoning with custom rubrics. Developed evaluation frameworks to score the LLM on crafted rubrics, achieving 27% across all tasks.

Data Scientist at Parlay Finance

May 1, 2025 - October 15, 2025

Engineered SBA loan classifiers, achieving 87% accuracy on ineligible applications to streamline loan approvals. Reduced false approvals by 22% using class rebalancing, early stopping, and learning rate tuning. Delivered stakeholder-ready reports translating SHAP results and KPI shifts into actionable updates for analytics teams.