Portfolio

Yash Prajapati

I build practical machine learning pipelines, real-time backend systems, and API-first products that turn messy technical problems into reliable software.

View projectsRead blogContact me

M.Sc. Data Science student at TUHH, based in Hamburg. Open to working student, internship, and research-driven engineering opportunities.

Current focusAvailable

Applied ML, MLOps, computer vision, distributed backend services, and developer tools.

PythonTensorFlowPyTorchJavaGolangKafkaDockerAWS

0%Accuracy on a Siamese Network facial recognition project.

0+Devices supported in a real-time Network Monitoring System.

0%Model accuracy improvement through feature engineering.

0%Latency reduction using async event-driven messaging.

Featured work

Portfolio case studies

These projects show the range I enjoy most: production-minded data science, high-throughput backend services, and tools that make technical workflows easier.

Network Monitoring System

Real-time infrastructure monitoring with Vert.x, Go plugins, ZeroMQ, Kafka, SQL storage, and a React dashboard.

Read case study

JavaGoKafkaReact

CodeCraftPro

Collaborative coding platform with real-time editing, workspace management, authentication, and Socket.IO synchronization.

Read case study

Node.jsReactSocket.IOMongoDB

One-Shot Face Recognition

CNN-based Siamese Network for facial recognition using contrastive learning, TensorFlow, OpenCV, and data augmentation.

Read case study

TensorFlowOpenCVDeep Learning

Experience

Engineering with measurable outcomes

My work spans ML pipelines, network monitoring platforms, REST APIs, and collaborative software development across research, public-sector, and product environments.

OOCL

Working Student - Data Analyst. Supporting data process automation and analytics at a global shipping company. Building Power BI dashboards and contributing to active digitalization projects in Hamburg.

BISAG-N

Software Developer - Data Science and ML. Built ML pipelines, tracked experiments with MLflow, and improved model reproducibility across geospatial datasets.

Motadata

Software Developer Intern. Built real-time NMS components with Java, Go, SQL databases, and async messaging for monitoring infrastructure at scale.

Writing

Modern engineering notes

I use the blog to explain how I approach systems, ML projects, and developer tools.

Real-time monitoring systems

How Vert.x, Go, Kafka, and ZeroMQ fit together in a scalable monitoring platform.

Read post

Siamese Networks

A practical walkthrough of one-shot facial recognition and contrastive learning.

Read post

MLOps for applied ML

Lessons from reproducible model training, validation, and experiment tracking.

Read post