Read Software Sushi articles on AI engineering, machine learning, software development, and production systems.
We deployed our multilingual support ticket classifier on one SageMaker ml.c5.xlarge and measured start-up time, real-time latency, batch throughput, and the cost per million tickets.
Read article22 articles in the journal
Use a furniture manufacturer’s forecasting decision to compare off-the-shelf AI, custom development, and a combined approach on business value and total cost.
Read article
How we built a private RAG pipeline on our own hardware: Docling, PII redaction, bge-m3, Qdrant hybrid search with an ACL filter, and a local model compared with a hosted one on 64 questions.
Read article
How we built a read-only MCP server for 43 AWS Marketplace AMIs in TypeScript, with live catalog data, the eight practices that shaped it, and what production still needs.
Read article
What an AI proof of concept should prove, how it differs from a prototype or pilot, what drives cost, and how to choose a PoC partner.
Read article
Give your AI agent a sandbox with rules. See how NVIDIA OpenShell fits with the model and agent, start a first sandbox on our AWS AMI, and watch a real task hit the boundary.
Read article
Send PHI to a healthcare LLM only after six questions: scope, provider, path, access, behavior, and operations.
Read article
A practical framework for evaluating RAG retrieval separately from generation with labeled queries, layered metrics, human review, and release gates.
Read article
A practical framework for assessing whether a healthcare AI application is ready to handle PHI across data flows, vendors, controls, testing, and operations.
Read article
How RAG and fine-tuning differ, what enterprises actually use, and how to choose based on changing data, evaluation, and total cost.
Read article
Why production ML models lose accuracy over time, how aging differs from drift, and what effective monitoring and intervention look like.
Read article
Why promising AI pilots stall before production, and what successful teams do differently across data, workflows, ownership, and integration.
Read article
A practical guide to planning AI projects that succeed — covering common pitfalls, MVP strategy, and the strategic groundwork most teams skip.
Read article
This post breaks down a common but costly mistake in AI forecasting: data leakage from random splitting — and shows how the right validation approach turned unreliable predictions into trusted business decisions.
Read article
A blog covering the basics of developing a chrome extension, as well as how to integrate one with the OpenAI API.
Read article
Fine-tune YOLO11n to detect hard hats on an Apple M4, export it to CoreML and ONNX, and run it frame by frame, with the code and every measured number.
Read article
Content moderation is a critical task for any online platform. The goal is to automatically identify and flag content that violates community guidelines, covering categories like hate speech, harassment, or self-harm.
Read article
Imagine having a virtual assistant that understands user preferences, context, and queries with precision—providing relevant and tailored responses in seconds. This is the promise of an AI-powered chatbot, a tool designed to make interactions more intuitive, efficient, and personalized using OpenAI API and Retrieval-Augmented Generation (RAG) frameworks such as LangChain.
Read article
In this first blog of a new three part series in OpenCV, we will be covering face detection using OpenCV.
Read article
A hands-on OpenCV guide to Sobel, Laplacian, and Canny edge detection, contours, morphological operations, template matching, Fourier and Hough transforms, and histogram equalization.
Read article
A practical OpenCV guide to color space conversion, blurring and filtering, thresholding, Otsu binarization, GrabCut, connected components, and watershed segmentation, with Python code.
Read article
A hands-on OpenCV guide to geometric transforms, image arithmetic, bitwise operations, masking, and color channels, with Python code and output for every technique.
Read article