Open to opportunities

Yogeshwar Shendye

@yogeshwarshendye

Message

I build production AI systems for computer vision and retrieval.

Canada

Message

What I'm looking for

I'm looking to build and deploy practical machine learning, computer vision, or LLM-driven systems where I can improve real-world workflows, work with production data, and collaborate across technical and domain teams.

I've built production AI systems at IBM, including an agentic RAG platform on Amazon Bedrock that improved retrieval accuracy by 35% and uptime by 25%.

I designed OpenSearch Serverless vector workflows that reduced context-retrieval latency from 250ms to 80ms, and automated ingestion and embedding refresh with Airflow to reduce manual intervention by 90%.

At Alentic Microscience, I developed microscopy computer vision pipelines and benchmarked CNN and Vision Transformer models for real-world imaging workflows. I processed 3,000+ images and reduced manual review time by 50% while addressing noisy annotations, class imbalance, and inconsistent sensor data.

My work spans medical-image segmentation, object detection, edge deployment, and LLM-driven data systems. I've also contributed to published research on platelet enumeration and AI-assisted instance segmentation in urology.

Experience

Work history, roles, and key accomplishments

Current

Data Engineer

Current

IBM

Apr 2025 - Present (1 year 5 months)

Built and owned a production-grade agentic RAG system using Amazon Bedrock, improving retrieval accuracy by 35% and system uptime by 25%. Designed and optimized vector search workflows with OpenSearch Serverless, reducing latency from 250ms to 80ms.

Airflow GitHub Actions CI CD RAG Vector Search Python

Computer Vision Researcher

Alentic Microscience

Apr 2023 - Aug 2024 (1 year 4 months)

Designed and optimized real-time computer vision pipelines for microscopy image acquisition systems. Built preprocessing and data ingestion pipelines for 3000+ microscopy images, reducing manual review time by 50%.

Computer Vision CNN Image Processing PyTorch TensorFlow Data Pipelines

Machine Learning Intern

PM AM IT Services

Sep 2022 - Dec 2022 (3 months)

Developed a real-time human pose estimation system using OpenVINO for video streaming analytics, achieving a 5% error rate under production constraints. Built and deployed an edge-optimized computer vision pipeline for volume estimation.

OpenVINO Computer Vision Docker Edge Computing Video Analytics Python

Data Science Trainee

MyVishwa

Nov 2021 - Mar 2022 (4 months)

Led end-to-end development of a real-time CNN-based object detection system for financial document verification. Designed preprocessing and feature extraction pipelines to improve classification performance.

CNN Object Detection data preprocessing Feature Extraction D3.Js Plotly Python