Suraj Subramanian Data Engineer

About Me Projects
Avatar

I'm currently a Data Engineer at ICEYE, actively contributing to the development and transition of SAR/Flood research prototypes to full-scale analytics products. My previous experience involved developing Customer Analytics solutions and creating efficient serverless pipelines at 247.ai. Throughout my academic journey in Computer Science and Engineering, I pursued three enriching internships in Data Science and Machine Learning.

My passion lies in the realms of Software and Infrastructure development, and I'm always on the lookout for fresh challenges and opportunities to learn and grow.

In the realm of Software, I've utilized Python to craft several production-ready applications. Lately, I've dived into the exciting world of Go and begun contributing to open-source projects I'm passionate about, a journey I look forward to expanding!

On the DevOps front, my journey began with serverless development on GCP at 247.ai. Over time, I've honed my skills in AWS (certified), Kubernetes (certifications - CKA, CKAD), and Terraform, empowering me to operationalize various applications.

Interested in working together or having a chat? Feel free to drop me a line on LinkedIn.

Personal Projects

Handwritten Equation Solver gif

Equation Solver

Application developed using ReactJS and Flask that recognizes and evaluates equations from an image you upload/write through sketchpad using OpenCV and a CNN model trained on the EMNIST dataset using Tensorflow

Check it out Source

Emotion Detection from Tom and Jerry videos gif

Emotion Detection from Tom and Jerry videos

Application recognises facial expressions from Tom and Jerry videos. Generated datasets using LabelImg and trained YOLO and VGG-19 models in PyTorch for Object Detection and Emotion Recognition into 4 categories - Happy, Angry, Sad, Surprised

Source

Sudoku Solver

Sudoku Solver

Application that solves sudoku puzzles from their photos. Puzzles are extracted using OpenCV, PyTesseract and a digit recognizer trained using PyTorch followed by a backtracking solver

Source

ML based Event predictor for small cells

Event predictor for Small Cells

Worked with a team of 5 members as part of the Samsung Prism initiative to build a pipeline that ships logs system logs from edge devices to a central server. Logs are used to forecast failures in edge devices using the Elasticsearch pipeline. Successfully predicted over 60% of the crashes with an acceptable number of False positives.

Publication