Loading…
Browse by type
Tutorial reads
Tutorial posts on Yomu, summarised newest first.
- 630 posts
- 20 companies
The Good Stuff
Paid picks we use ourselves. Links carry a referral code — a subscription through one earns us a commission at no extra cost to you.
Tutorial posts
Building a .NET API Gateway with ClickHouse and Aspire
The tutorial builds a .NET distributed application around a YARP API gateway, two backend APIs, an analytics API, a Blazor dashboard, a load generator, and ClickHouse. Aspire’s AppHost coordinates these resources, injects connection strings and service endpoints, orders startup with dependency health checks, and exposes selected services through stable or external endpoints. The gateway records proxied requests in ClickHouse while emitting OpenTelemetry metrics for request counts, duration, and backend failures. A bounded queue and background ingestion keep request logging off the hot path, while a narrow schema and materialized view precompute minute-level aggregates for volume, p95 latency, error rates, and route comparisons. Aspire supplies live telemetry, whereas ClickHouse retains request history for analytical queries and drill-down into slow requests.
Profiling in PyTorch (Part 1): A Beginner's Guide to torch.profiler
The post introduces torch.profiler through a beginner-oriented walkthrough of profiling a matrix multiplication and addition, explaining how to configure the profiler and interpret its outputs. It uses a small function with torch.matmul and torch.add, annotates it with record_function, captures CPU and CUDA activities after warm-up iterations, and exports a profiler table plus a Chrome trace. The table summarizes event counts and CPU or GPU time, while the trace shows execution order, call hierarchy, CPU-GPU gaps, kernel launches, and overlap. Comparing 64x64 and 4096x4096 matrices demonstrates a shift from overhead-bound behavior to compute-bound behavior; the guide also covers dispatch chains, hidden device-to-device copies, profiler buffer activity, and torch.compile's runtime and caching behavior.
6 Tips for Accessible Emails
The Email Markup Consortium’s 2026 report analyzed 376,348 emails and found that only eight passed every accessibility check, while nearly all failures included Serious or Critical issues. The post focuses on six largely mechanical improvements: setting lang and dir, marking layout tables role=“presentation”, creating a heading hierarchy, writing descriptive link and image text, adding a specific title element, and meeting a 4.5:1 color-contrast ratio. It gives HTML examples, notes that clients may strip attributes from the html element, and distinguishes meaningful images from decorative ones that need empty alt text. It also describes updated React Email defaults for language, direction, image alt text, Markdown tables, and titles, while leaving content judgments such as headings, links, and contrast to authors.
The ultimate guide to knowledge management for your Service Agent
AI-driven customer service depends on knowledge management because a Service Agent needs accurate product, feature, policy, and troubleshooting information to resolve queries end to end. The guide divides the work into two phases: building a knowledge base, and maintaining, optimizing, and scaling existing content. It recommends using support conversations and tools such as Fin Recommendations to identify gaps, prioritize FAQs, onboarding guides, troubleshooting material, and other resources, then centralize content and control what the Agent can use. Ongoing practices include scheduled reviews, a request log, launch-planning updates, consistent terminology, audience targeting, and contact details for escalation. The stated conclusion is that knowledge management should be a continuous core service function, with improvements producing more resolutions, better data, and further opportunities to refine support.
Reachy Mini goes fully local
Reachy Mini’s conversation app can now run its speech stack locally instead of sending audio to a server, avoiding cloud, API keys, and data leaving the machine. The tutorial uses speech-to-speech as a cascaded VAD → STT → LLM → TTS pipeline exposed through a Realtime API-compatible /v1/realtime WebSocket, with Silero VAD v5, Parakeet-TDT 0.6B v3, Qwen3-TTS, and a selectable LLM backend. It shows llama.cpp serving Gemma 4 E4B with parallel slots, 64k context, flash attention, and full sliding-window attention, then covers vLLM, MLX, Transformers, hosted Responses API endpoints, and network connection to a wireless robot. Once the backend and app are connected, conversations run through the local loop, while individual stages can be exchanged to trade latency, quality, language coverage, and hardware choices.
Observability for any agent, anywhere: Production-ready tracing with OpenTelemetry & Unity Catalog on Databricks
Databricks supports writing OpenTelemetry (OTel) traces directly to Unity Catalog, where real-time telemetry is stored in Delta tables for governed analytics and retention. AI traces capture prompts, tool calls, responses, latency, and execution paths, enabling debugging, evaluation, and monitoring, while lakehouse storage also allows SQL queries, dashboards, joins with business data, and PII controls. The managed, serverless ingestion layer uses Zerobus Ingest to accept OTLP over gRPC from collectors and REST integrations, streaming spans, logs, and metrics to Unity Catalog without intermediate message buses. A LangGraph support manager assistant demonstrates instrumentation with mlflow.langchain.autolog() and an @MLflow.trace root span, while Genie is invoked through MCP for data-driven questions. The resulting traces can be searched in MLflow, evaluated at scale, and monitored continuously, with FAQ details stating 200 QPS starting throughput, no storage limit, and Unity Catalog governance options for access control, masking, and filtering.
Native random sampling in ClickHouse
ClickHouse native random sampling enables aggregate queries to execute against a representative fraction of data when querying full datasets is too slow. To maintain sampling accuracy, tables require a sample key derived from high-cardinality columns using hash functions like sipHash64 to ensure an even distribution. The sample key must be included in the primary key ORDER BY expression, preferably at the front to let ClickHouse quickly skip index granules. Queries can sample data deterministically by specifying a fractional ratio or a minimum row threshold, scaling aggregate counts using the virtual column _sample_factor. In benchmark tests on a 30-million-row dataset, sampling 10% of records reduced query latency by approximately 80% while retaining approximate metrics suitable for exploratory analysis.
Six Steps to Improve Your Sender Reputation
Sender reputation is the largely private assessment that inbox service providers such as Gmail, Outlook, and Yahoo make before deciding where to deliver email. It reflects infrastructure compliance, sending history and consistency, list hygiene, mail frequency, and recipient engagement, with providers increasingly emphasizing domain reputation over IP reputation. The guide recommends six recovery steps: audit SPF, DKIM, and DMARC; warm up new domains gradually; identify provider-specific blocks or blocklists; clean and segment lists; make unsubscribing easy; and send valuable, mobile-friendly content. It also advises using double opt-in, monitoring engagement, bounce and spam responses, throttling, and DMARC reports, while tightening policies or requesting provider reviews after fixes. Because providers continually reevaluate senders, reputation management requires repeated observation, testing, and adjustment rather than a one-time repair.
How to Build Real-Time Fraud Detection using Spark Real-Time Mode and Lakebase
Card fraud operates in seconds, while organizations need to score and block suspicious transactions before settlement without adding a separate streaming stack. This Solution Accelerator presents an open-source, end-to-end implementation on Databricks, covering transaction ingestion, Spark Real-Time Mode processing, fraud scoring, Lakebase feature serving, and a Databricks Apps monitoring dashboard. RTM processes streaming workloads in milliseconds and targets sub-300ms latency, while Lakebase supplies a managed, serverless Postgres serving layer with sub-millisecond reads. The design keeps batch ETL, training, streaming, and governance within the same platform. Deployment progresses from a synthetic rate-source quick start to Kafka-based parsing, stateful per-card velocity tracking with TTL, enrichment, weighted scoring, routing, and optional MLflow-trained RandomForest scoring, followed by a Streamlit app for live decisions.
How to safeguard AI workloads with Unity AI Gateway Guardrails
Unity AI Gateway’s beta LLM Guardrails extend earlier Gateway guardrails with LLM-powered evaluation, pre-built protections, and tunable custom rules for AI security, compliance, and sensitive-data protection. An Acme marketing scenario maps policies to PII redaction on inputs, jailbreak and prompt-injection blocking, unsafe-content blocking on outputs, and a custom rule that blocks competitor references. Teams configure these guardrails on an endpoint, optionally use Log mode, select evaluator endpoints, and test representative requests while tracking activity in inference tables. Built-in controls behave as expected in the examples, while the custom rule initially performs unreliably until a more specific prompt and gpt-5-4-mini evaluator improve triggering without degrading other tests. Inference tables record request outcomes, evaluator calls, latency, statuses, token counts, and shared request_id values for auditing and refinement.
18 Social Media Holidays to Celebrate This June
June’s social media calendar presents 18 observances, combining lighthearted dates such as National Donut Day and National Take Your Dog to Work Day with causes including ocean conservation and blood donation. It positions the holidays as prompts for brands to engage online communities through polls, carousels, infographics, videos, challenges, and behind-the-scenes posts. Suggested hashtags and content ideas range from a childhood-memory question and donut personality poll to environmental restoration visuals, ocean-health graphics, blood-donation myth busting, and a collaborative summer playlist. June 21 carries several entries—Father’s Day, the first day of summer, International Yoga Day, National Selfie Day, World Music Day, and National Take Your Dog to Work Day—while later guidance covers sunglasses and Social Media Day content creation.
Unlocking asynchronicity in continuous batching
Continuous batching improves GPU utilization with tightly packed batches, but its synchronous loop makes the CPU and GPU take turns, leaving the GPU idle while the CPU prepares each next batch. Profiling 8K-token generation with a batch size of 32 and an 8B model recorded 300.6 seconds, with the GPU idle for 24.0% of runtime. The post builds asynchronous batching by overlapping CPU preparation with GPU computation, using non-default CUDA streams, non-blocking transfers, and synchronization events to coordinate dependencies. In the same experiment, GPU activity rose to 99.4% and total generation time fell to 234.5 seconds, a 22% speedup, without new kernels or model changes while maintaining model accuracy.
The ultimate guide to knowledge management for your Sales Agent
AI-powered inbound sales depends on knowledge management: accurate, current information about products, pricing, plans, qualification, and routing gives a Sales Agent the context to answer questions and guide buyers. The guide describes knowledge management as creating, organizing, sharing, and maintaining business knowledge, while emphasizing content such as product FAQs, pricing overviews, competitor battlecards, case studies, and internal sales materials. It recommends finding gaps through recurring questions, buyer objections, search data, and disengaged conversations; then writing machine-friendly content with simple language, consistent terminology, explanatory context, scannable formatting, and text accompanying images or video. Connecting the Agent to CRM and enrichment systems keeps data and context connected, while ongoing reviews help content evolve from conversation insights and reduce repetitive work for sales teams. It presents knowledge management as a continuous sales function whose cumulative improvements can support faster responses, better qualification, and more focused human selling.
Are you winning in AI search? Do this self-assessment.
AI search is reshaping consumer discovery, but the post says most brands do not know what AI platforms say about them. It presents a four-step assessment: choose one country, define a focused category, identify roughly 100 consumer questions, and test whether the brand appears in answers from ChatGPT, Gemini, Perplexity, and Google AI Overviews. Search Intelligence platforms track queries, mentions, recommendations, citations, and content gaps across search and AI tools. The post recommends answering real queries with dedicated content, earning coverage from accessible independent sources, participating helpfully in online conversations, and monitoring AI answers monthly because they change. Its examples report a 25% reduction in Orange call-center contacts, 50% lower Campari costs, and Aromazon reaching €40 million in annual revenue.
How to Build Low-Code API Integrations for Enterprise Apps Using Okta
Okta’s API Integration Actions let independent software vendors build Provisioning, Entitlements, and Universal Logout integrations in the Okta Integration Network using Workflows, Okta’s low-code builder. The approach maps existing APIs directly to Okta action contracts, allowing developers to avoid building and maintaining a SCIM server while translating API calls into formats Okta can consume. The guide walks through creating an OIN integration, configuring tenant and authentication mappings, adding API Integration Action components, building and validating flows, and submitting them through the OIN Wizard. It also explains how to create and update an org-only test instance before review, noting that submission changes do not automatically update an existing test instance for security reasons.
Learnings from init to deploy
An agent was used to build a transcription app for a team from initialization through deployment with Stripe Projects, OpenRouter, and Vercel. The app sends video-generated subtitles through an OpenRouter agent with a detailed system prompt and Stripe documentation tool calls, producing a formatted transcription whose speaker and service names are corrected. The author recommends treating agent-generated plans like design documents, verifying provider behavior, and using streaming for long transcripts because Vercel serverless limits made non-streaming responses prone to 504s for a roughly 1,500-line transcript. Testing led to removing an unreliable free-model tier, defaulting to openrouter/auto, and choosing Claude Opus 4.5-thinking over Claude Sonnet 4.5; shared agent files were committed, and a separate review agent was used before human review.
Database modeling: a practical guide to techniques and best practices
Database modeling defines the structure, relationships and constraints that keep systems coherent across transactional, analytical and flexible workloads. The guide presents a three-phase design process: conceptual modeling captures entities and relationships, logical modeling adds attributes, keys, cardinality and normalization, and physical design translates that model into tables, indexes, constraints, partitioning and other platform-specific choices. It compares relational, document/NoSQL, dimensional, hierarchical and network models, linking each to data shape, workload and consistency requirements. It recommends explicit keys, consistent naming, appropriate normalization, query validation and version-controlled DDL, while warning against premature physical design, missing foreign keys, unclear NULL handling and poorly balanced schemas. It concludes that Databricks Lakebase combines a transactional database engine with the Databricks Lakehouse Architecture so one model can support operational and analytical workloads.
From init to deploy: building with agents and Stripe Projects
The post describes an internal agentic transcription app built to reduce the manual effort required to correct auto-generated subtitles for Stripe Developers videos. The app uses a Next.js interface with transcript, chat, and corrected-output panels, while Stripe Projects CLI links Vercel and OpenRouter and populates required environment variables. An LLM client supports selectable models, iterative instructions, and a transcription agent whose system prompt emphasizes preserving transcript structure; the agent can also use Stripe documentation search through MCP. In a benchmark on a 771-line meetup script, Gemini 2.5 Pro reached 99% accuracy, while the tested models reduced editing from more than 30 minutes to just over two minutes of generated edits plus a human check. The reported overall reduction was 4x, with testing costs totaling $2.54.
Comparing ClickHouse versions with clickhousectl
Evaluating query performance differences across ClickHouse versions previously required managing Docker containers or manually downloading historical binaries from GitHub releases. The clickhousectl command-line interface simplifies this process by downloading specific ClickHouse binaries, running multiple background server instances concurrently with automatic port assignment, and executing queries across local environments. Users can copy table schemas between local instances by piping raw output from SHOW CREATE TABLE queries and replicate datasets using ClickHouse's remote table function. In a benchmark comparing ClickHouse 25.12 and 26.3 on an S3-backed Parquet dataset, version 26.3 introduced a Parquet metadata cache that reduced repeat query times from roughly 9 seconds to 1–2 seconds. System tables like system.server_settings and system.metrics enable operators to inspect cache configurations, eviction policies, and memory consumption directly.
How to build scalable web apps with OpenAI's Privacy Filter
The article presents three web applications built around OpenAI’s open-source Privacy Filter, a PII detector that labels eight categories in one forward pass over a 128k-token context. Document Privacy Explorer extracts PDF or DOCX text and returns category-labeled spans, while Image Anonymizer combines Tesseract word boxes with character offsets to produce editable pixel redaction rectangles. SmartRedact Paste replaces detected spans with category placeholders, stores public and token-gated views, and evicts expired pastes in a daemon thread. Across the apps, gradio.Server’s @server.api routes model work through Gradio’s queue and gradio_client, while plain FastAPI routes serve custom HTML, views, and lightweight lookups. The implementations keep browser-side filtering, canvas edits, and PNG export out of server round-trips, and the paste service is about 200 lines including storage.
Top topics
- Architecture834
- Performance817
- LLMs673
- Machine Learning622
- Open Source568
- AI Agents564
- AI483
- Developer Experience448
Weekly digest
The best engineering reads, once a week.