51eshare's Tech Blog
This site is dedicated to sharing IT-related technologies.
Building a Distributed Keras Training System on Hadoop YARN for Terabyte-Scale Audio Emotion Analysis Building a Distributed Keras Training System on Hadoop YARN for Terabyte-Scale Audio Emotion Analysis
The prototype worked flawlessly. A Keras model, a convolutional recurrent neural network, was achieving respectable accu
2023-10-27 MLOps
Emotion Hadoop Keras PySpark YARN Distributed Training
Integrating a Swift Client with a Legacy MyBatis Backend Using an Azure-Hosted SciPy Computational Service Integrating a Swift Client with a Legacy MyBatis Backend Using an Azure-Hosted SciPy Computational Service
The initial request was seemingly straightforward: enable a new iOS application for our portfolio managers to perform re
2023-10-27 Architecture
Microservices Azure Swift SciPy MyBatis Turbopack Integration
Implementing a Fault-Tolerant Stateful Aggregation Layer in Dask with LevelDB Backends Implementing a Fault-Tolerant Stateful Aggregation Layer in Dask with LevelDB Backends
The core challenge with large-scale, long-running stream aggregations isn’t the computation itself, but the management o
2023-10-27 Data Engineering
Distributed Systems Dask LevelDB BASE Python
Injecting Custom Business Tags into SkyWalking Traces via a Kong Lua Plugin for ASP.NET Core Services Injecting Custom Business Tags into SkyWalking Traces via a Kong Lua Plugin for ASP.NET Core Services
The observability stack was reporting clean traces, but they felt hollow. We had our services instrumented with the SkyW
2023-10-27 Observability
Distributed Tracing Kong ASP.NET Core SkyWalking Lua
Orchestrating Containerized Android Emulators on AWS EKS for Integration Testing with Vitest Orchestrating Containerized Android Emulators on AWS EKS for Integration Testing with Vitest
The shared, static pool of Android emulators had become the primary bottleneck in our CI pipeline. Contention was consta
2023-10-27 DevOps
CI/CD Vitest Kubernetes Android Go AWS EKS
Constructing a Real-Time NLP Enrichment Pipeline with Batch Processing and Multi-Tier Caching Constructing a Real-Time NLP Enrichment Pipeline with Batch Processing and Multi-Tier Caching
The initial system design was straightforward and, in hindsight, naive. An API endpoint received a block of text, passed
2023-10-27 Data Engineering
Python Rollup Memcached spaCy Snowflake asyncio
6 / 52