Nithin E.

Key Skills

Software

OpenCV AI Kit (OAK)

Other

Top Subject Matter

Semiconductor manufacturing (MES alarms, equipment events)

Semiconductor wafer inspection (defect detection/segmentation)

Industrial welding (video-based object detection and tracking)

Top Data Types

Text

Image

Video

Document

Top Task Types

Segmentation

Object Detection

Classification

Bounding Box

Tracking

Freelancer Overview

Algorithm Developer — MES Data Analytics (alarm/event NLP clustering and correlation). Brings 11+ years of professional experience across complex professional workflows, research, and quality-focused execution. Core strengths include Internal, Proprietary Tooling, and OpenCV AI Kit (OAK). Education includes Master of Science, Indian Institute of Technology Madras (2022) and Bachelor of Engineering, Vasavi College of Engineering (2016). AI-training focus includes data types such as Text, Image, and Medical and labeling workflows including Entity (NER), Segmentation, and Object Detection.

Labeling Experience

Algorithm Developer — MES Data Analytics (alarm/event NLP clustering and correlation)

Text

Developed Transformer-based NLP analytics to cluster and correlate semiconductor MES alarm and event text across multiple sources for faster root-cause identification. Correlated initiating alarms with process run data to map relevant sensor parameters within event windows and highlight recurring failure patterns. Used the outputs to support yield-loss prevention and improved tool uptime during wafer fabrication. • Input: alarm/event text streams from MES and equipment systems • Labeling/annotation goal: deriving structured alarm groups, initiating event identification, and associated process/sensor attributes • Output artifacts: correlated alarm sequences, anomaly signatures, and reporting for engineers • Impact: reduced repetitive excursions and accelerated troubleshooting of yield-impacting tool issues

2024 - Present

Associate Research Engineer - Digital Systems - GKN Aerospace Engine Systems

ImageSegmentationClassification

Served as an associate research engineer developing digital systems and applied machine learning for manufacturing process monitoring. Designed real-time computer vision for tracking welding wire behavior and stability, and created models to detect rare anomalies in process video. Produced segmentation, classification, and visualization components using modern ML toolchains and contributed to technology documentation aligned with research readiness processes. • Developed YOLOv8-based real-time wire tracking and process stability monitoring • Built convolution autoencoders to detect rare anomalies during manufacturing video capture • Implemented TransUNet-based segmentation for melt pool separation from wire • Used XGBoost classification and Plotly dashboards for production-grade monitoring by engineers

2023 - 2024

Associate Research Engineer — Digital Systems (unsupervised anomaly detection)

VideoClassification

Developed unsupervised convolutional autoencoder methods to detect rare anomalies in manufacturing videos. Focused on learning normal video patterns and flagging deviations during the manufacturing process. Used results to identify unusual events that could indicate quality issues. • Input: manufacturing video sequences • Labeling/annotation goal: anomaly discovery and classification without exhaustive manual labeling • Output artifacts: anomaly scores or flags for operational review • Impact: faster detection of rare abnormal occurrences in production footage

2023 - 2024

Associate Research Engineer — Digital Systems (wire detection/tracking with YOLOv8)

Object Detection

Implemented real-time wire detection and tracking for welding process monitoring using YOLOv8. Tracked wire position within the melt pool and evaluated process stability during manufacturing. Correlated unstable welding regions with pore locations derived from CT scan analysis to investigate root causes of internal component defects. • Input: video frames from welding process and CT-derived imagery for correlation • Labeling/annotation goal: generating object detections and tracking targets for wire/melt-pool features • Output artifacts: stable/unstable monitoring indicators for engineers • Impact: improved manufacturing quality by supporting defect root-cause investigations

2023 - 2024

Engineer - Image Processing - Transasia Bio-Medicals Research And Development

ImageSegmentation

Worked as an image processing engineer applying deep learning and classical computer vision for medical device perception tasks. Implemented disparity estimation for a phlebotomy-based robot to estimate depth of hand and vein surfaces using Unimatch. Built vein segmentation workflows for robotic venepuncture and developed supporting software components for operational use. • Predicted 3D disparity maps using Unimatch-based deep learning for depth estimation • Implemented YOLOv5 for antecubital fossa detection using labeled images • Designed disparity estimation using Semi-Global Block Matching to compute depth at ROI • Performed vein segmentation using image binarization, Hessian processing, and wavelet transforms and built basic PySimpleGUI software

2023 - 2023

Show more

Education

Work History