Open to opportunities

Dushyant Shrivastava

@dushyantshrivastava

Message

I build low-latency computer vision systems for real-time defect detection.

India

Message

What I'm looking for

I'm looking to build and deploy real-time computer vision and machine learning systems, especially object-detection solutions where I can improve model accuracy, inference latency, and production reliability.

I've built and shipped YOLOv8-based defect detection and classification systems at AN Nextgen Innovation, taking models from dataset preparation and annotation through training, evaluation, optimization, and deployment.

I improved inference latency from 450ms to 85ms using ONNX and TensorRT INT8 quantization, while building a FastAPI, CustomTkinter, and SQLite dashboard that processes 1,000+ events per second with live model-result visualization. My work includes investigating misclassified edge cases, improving pipeline quality, and validating models against labeled test sets.

Previously, I deployed PyTorch detection models for AP Technologies and built custom YOLOv8 and OpenCV inspection models at Ethara AI, achieving 91% recall and reducing manual QC time by 35%.

Experience

Work history, roles, and key accomplishments

Current

AI Engineer (R&D) - Computer Vision

Current

AN Nextgen Innovation Pvt. Ltd.

Mar 2026 - Present (5 months)

Built and trained a YOLOv8 object-detection/classification model for defect QC, covering the full pipeline from dataset prep to deployment. Profiled and tuned inference with ONNX and TensorRT, reducing latency from 450ms to 85ms, and developed a real-time dashboard processing 1000+ events/sec.

Python PyTorch YOLOv8 OpenCV Onnx TensorRT fastAPI SQLite & Model Tuning

Freelance AI/ML Engineer

AP Technologies

Nov 2025 - Feb 2026 (3 months)

Trained and deployed PyTorch object-detection models as ONNX and TensorRT, achieving a 3.2x inference speedup. Reduced model footprint by 60% via DLL-based integration for deployment across legacy systems.

PyTorch Onnx TensorRT Deployment C++

AI Engineer Intern - Computer Vision

Ethara AI

Aug 2025 - Oct 2025 (2 months)

Built and trained custom YOLOv8 and OpenCV object-detection models for defect inspection, preparing and annotating datasets. Ran evaluation jobs to generate performance reports, improving model iterations and cutting manual QC time by 35%.

Python PyTorch YOLOv8 OpenCV Data Annotation Model Evaluation Precision Recall

Data Analytics Intern

Unified Mentor

May 2025 - Jul 2025 (2 months)

Built data ingestion and cleaning pipelines using Python, Pandas, and SQL on 50,000+ records. Created 5 Power BI dashboards for reporting and experiment documentation.

Python Pandas SQL Power BI Data Cleaning Data Pipelines