Loading…
Browse by type
Tutorial reads
Tutorial posts on Yomu, summarised newest first.
- 615 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
End-to-End RAG Workflow: How Retrieval Augmented Generation Works
Retrieval Augmented Generation (RAG) connects a large language model to external knowledge at inference time, addressing outdated training data and access to proprietary or real-time information. The workflow ingests and normalizes sources, splits documents into chunks, embeds them in a vector store, retrieves context, assembles a prompt, and generates an answer. Semantic search can be combined with BM25 keyword search through reciprocal rank fusion, while reranking can improve precision; the same embedding model must be used during ingestion and querying. The guide presents evaluation and deployment considerations, including separate measurement of retrieval precision and generation faithfulness, versioning, monitoring, and containerized components. It identifies poor retrieval as the most common failure mode and explains that RAG reduces, but does not eliminate, hallucinations.
Experimenting with the proposed Cross-Origin Storage API in Transformers.js
Transformers.js browser apps currently redownload and recache identical AI model and WebAssembly resources when they are requested from different origins. Chrome's cache uses a Network Isolation Key composed of the top-level site and current-frame site, so matching CDN URLs do not produce cache hits; the toy example incurs 177 MB of duplicate model storage, while a shared ONNX Runtime Wasm file is 4,733 kB. The proposed Cross-Origin Storage (COS) API addresses this by identifying large files with cryptographic hashes and sharing them across opted-in origins, with a polyfill extension enabling experimentation before native browser support. In the extended Transformers.js flow, a COS lookup can serve resources in milliseconds, while a miss downloads and stores them, and ModelRegistry.is_pipeline_cached() can help select among models already cached.
Build Cross-Language Multi-Agent Team with Google’s Agent Development Kit and A2A
The post presents a contract-compliance pipeline in which a Python agent uses Gemini to extract contract terms and a Go service applies deterministic corporate-policy checks. It connects the services with the Agent2Agent (A2A) protocol and uses Google's Agent Development Kit (ADK), including RemoteA2aAgent, to expose the remote validator as a local sub-agent in a SequentialAgent. A2A supplies discovery through Agent Cards at /.well-known/agent.json, JSON-RPC 2.0 messaging through message/send, typed Message Parts, and task states such as submitted, working, completed, and failed. Pipeline state is shared through ADK's ToolContext.state, with failures transitioning to MANUAL_REVIEW for human legal review when the Go service is unavailable. The accompanying open-source application demonstrates live handoffs, policy verdicts, and simulated network faults, while the conclusion emphasizes language independence, decoupling, resilience, and deterministic auditing.
Conversation design: How to make your AI Agent communicate like your team
AI agents need deliberate conversation design to communicate effectively with customers, covering tone, response structure, handoffs, interaction flow, and response quality. The post argues that without an explicit owner, an agent may over-explain, sound flat during frustration, or escalate too late, undermining trust and prompting customers to seek human help. It recommends defining a concise voice reference, adapting register to the situation, matching detail to the question, supporting customer-led topic changes, designing informative handoffs and follow-ups, and keeping behavioral instructions separate from knowledge-base content. Teams can begin without a dedicated hire by naming an owner, selecting a poorly performing conversation type, auditing instructions, and fixing the worst handoff first. At Fin, an A/B test raised CSAT from 72.8% to 78.4%.
Data Pipeline Best Practices: Architecture, Modern Pipelines, and Deployment
Modern data pipelines move raw data from source systems through transformation and storage to downstream consumers, with batch, streaming, or hybrid operation selected according to latency needs. The guidance places service-level agreements first, defining acceptable latency, uptime, and error rates before architecture choices, and compares ETL, ELT, and zero-ETL patterns. It recommends immutable raw landing zones, CDC for transactional sources, schema registries, event replay, incremental loads, and idempotent writes using techniques such as MERGE operations. Performance and reliability practices include partitioning, clustering, data-quality checks, observability, automated CI/CD testing, and cost controls such as autoscaling and storage tiering. Adoption guidance advises piloting a representative pipeline, measuring telemetry and cost, then iterating through regular architecture reviews.
Building a SQL ETL Pipeline: The Complete Guide for Data Engineers
SQL ETL pipelines are presented as repeatable workflows that extract data from sources, transform it, and load it into warehouses, lakes, or lakehouses for analysis and machine-learning use. The guide addresses source connectivity, extraction patterns, transformation logic, loading targets, governance, performance, testing, and operational design, while contrasting ETL with ELT and broader data pipelines. It explains that SQL can serve as the primary implementation language for transformations and load operations, with techniques including JOIN and GROUP BY, window functions, MERGE upserts, and deduplication with ROW_NUMBER() or DISTINCT. It also covers full versus incremental extraction, batch and streaming needs, schema-on-write versus schema-on-read, and layered validation using row counts, checksums, business rules, and schema-drift monitoring.
Design Beautiful Dashboards in AI/BI
This guide presents dashboard design as a way to apply brand identity consistently while helping viewers interpret business data and trust the result. It recommends choosing a 3-, 4-, or 6-column subdivision of AI/BI’s 12-column grid according to audience needs, then using F- or Z-pattern hierarchy and unequal widget sizes to prioritize information. The visual system combines readable sans-serif or local fonts, high-contrast text, neutral backgrounds, a 60-30-10 color allocation, and distinct interactive and visualization palettes. For the Wanderbricks example, the final five-color palette is #15AFDD, #2375A8, #52A870, #C85070, and #C89930, tested in light and dark modes and with color-blindness simulations; workspace themes extend the resulting styling across dashboards.
From the Hugging Face Hub to robot hardware with Strands Agents and LeRobot
Strands Robots integrates Hugging Face’s LeRobot stack with Strands Agents to connect dataset recording, policy execution, simulation, hardware deployment, and fleet coordination in one agent loop. Its integration keeps LeRobot responsible for hardware drivers, calibration, recording, and dataset handling, while Strands supplies AgentTools, a shared Robot abstraction, policy providers, and a Zenoh peer mesh; simulation and hardware use the same LeRobotDataset format. The walkthrough records demonstrations in MuJoCo, optionally pushes them to the Hub, runs policies through a common interface, and switches from simulation to a physical SO-101 with mode="real". The companion script and notebook run the default simulation workflow on a laptop without hardware, a GPU, or Hugging Face credentials, although the default Mock policy produces structurally valid but behaviorally unhelpful random demonstrations.
Build Your Own Transaction Foundation Model for Financial Intelligence
Production financial intelligence systems frequently rely on hand-engineered tabular features and rule sets that are brittle, expensive to maintain, and blind to historical sequential patterns. NVIDIA demonstrates an accelerated reference pipeline that replaces standard BPE tokenization with a GPU-based domain tokenizer, converting raw transactions into semantic tokens with an 8,192-token context window. Using the NeMo AutoModel library, a compact 29-million parameter decoder-only transformer is pretrained from scratch on unlabeled transaction sequences using causal language modeling. Learned sequence embeddings are extracted, compressed using PCA, and concatenated with raw tabular features to train a downstream GPU-accelerated XGBoost fraud detection model. On the IBM TabFormer benchmark dataset, this combined approach achieves a 41.76% lift in Average Precision over the baseline model relying solely on raw tabular features.
From hobby to real world: integrating the hard stuff with agents
The post documents building a production-oriented leaderboard for open-source Stripe Projects applications, where users submit projects, share links, and vote. Unlike an earlier transcription app, this project required authentication, multiple database tables, hosting, and controls limiting unauthenticated users to one vote per project. The agent initially selected Supabase for database and auth and Vercel for hosting, then added Upstash Redis after an in-memory rate limiter proved unsuitable for serverless enforcement. Supabase's two-emails-per-hour authentication limit prompted a human-directed migration to Clerk, after the agent proposed an unsuitable Twilio solution. The resulting MVP used five resources across four providers, took more than a day, and received separate security and performance audits before review.
How to Rent a Car in Grand Rapids, Michigan
Renting a car in Grand Rapids can provide flexible transportation around the city and for trips to nearby areas, including for travelers arriving at Gerald R. Ford International Airport. Uber Rent lets users compare vehicles from rental companies in the Uber app or online, with choices including economy cars and SUVs; available suppliers may include Budget, Hertz, and Avis, depending on location. The booking process covers entering pickup and drop-off details, reserving a vehicle, collecting it with a matching driver’s license and physical credit card, and returning it to the listed location. Renters must be at least 18, while a Young Driver Fee may apply under 25; personal insurance is not required, though supplier-provided options are available, and additional conditions may appear in the reservation terms.
How to Rent a Car in Fayetteville, North Carolina
Renting a car in Fayetteville, North Carolina, can provide mobility for airport arrivals, city-center work, local errands, and trips to surrounding areas. The guide describes using Uber Rent through the Uber app or online to book available sedans, SUVs, and other models from rental companies such as Budget, Hertz, and Avis, with availability varying by location. It recommends matching vehicle type and rental duration to the trip, while reviewing rates, mileage, and additional fees, then outlines reservation, pickup, payment, and drop-off steps. Renters must be at least 18, present a valid license matching the reservation, and bring a physical credit card in the primary driver's name; under-25 renters may pay a Young Driver Fee, and rental companies can offer insurance and protection plans.
How to Rent a Car in Spokane, Washington
Renting a car in Spokane can provide flexibility for city travel, work visits, longer stays, and drives to nearby destinations. Uber Rent lets travelers browse, compare, and book vehicles from rental companies in the Uber app or online, with options including compact cars, SUVs, and same-day or advance reservations. The booking flow covers entering pickup and drop-off details, completing payment either in advance or at the rental counter, collecting the vehicle with a valid driver’s license and physical card, and returning it to the stated location. The source says renters must be at least 18, while a Young Driver Fee may apply under 25; personal insurance is not required, though suppliers may offer protection plans, and requirements can vary by vehicle and supplier.
How to Rent a Car in Peoria, Arizona
Renting a car in Peoria, Arizona, is presented as an option for local transportation, regional trips, and travel from Phoenix Sky Harbor International Airport (PHX). Uber Rent lets users compare and book vehicles from participating rental companies, including Budget, Hertz, and Avis, through the Uber app or online, although availability varies by location. The booking process covers entering pickup and dropoff details, reserving a vehicle, collecting it with a driver’s license and payment method, and returning it to the listed location; reservation details appear in the app and confirmation email. The source advises matching vehicle type and rental duration to the trip, reviewing base rates and extra fees, and checking requirements such as a minimum age of 18, a physical credit card in the primary driver’s name, and supplier-specific conditions.
How to Rent a Car in Bakersfield, California
Renting a car in Bakersfield can provide flexibility for city travel, trips beyond the city, airport arrivals through Meadows Field Airport (BFL), business stays, or extended visits. The guide describes using Uber Rent to browse, compare, and book vehicles from rental companies, with options ranging from sedans to SUVs and reservations available through the Uber app or online. It outlines choosing a vehicle based on trip purpose, space, rental duration, and budget, then covers reserving, picking up, and returning the car while keeping booking details in the app and confirmation email. Requirements include being at least 18, presenting a matching valid driver’s license and physical credit card, and reviewing supplier-specific conditions; personal car insurance is not required, though protection plans may be offered at pickup.
Talk to all your data, wherever it lives
Lakehouse Federation addresses the challenge of reasoning across enterprise data spread among AWS Glue, Snowflake, Oracle, BigQuery, Postgres, and legacy formats without first migrating it. By connecting external sources in place and syncing their metadata into Unity Catalog, Databricks applies shared permissions, lineage, and access controls while preserving source data. Federated comments and descriptions give Genie schema context, while Unity Catalog Semantics lets teams define governed metrics such as ROI once for consistent use across Genie, dashboards, and notebooks. The example connects an AWS Glue marketing database, carries its metadata, defines an ROI metric view, and asks Genie which campaigns led ROI last quarter. The post reports an immediate, accurate answer from live Glue data, and describes planned richer semantics, broader federation, and possible performance gains from managed tables.
Profiling in PyTorch (Part 2): From nn.Linear to a Fused MLP
Part 2 of the series examines how PyTorch profiler traces change from a single bias-enabled nn.Linear to a three-layer MLP with an intervening activation, using scripts run on an NVIDIA A100-SXM4-80GB GPU. It explains that nn.Linear represents y = x @ w.T + b, dispatches aten::addmm, and folds bias addition into the cuBLAS GEMM epilogue, while aten::t only rewrites weight strides and launches no GPU kernel. For the MLP, eager execution uses three GEMMs plus separate activation and multiplication kernels, whereas torch.compile fuses the latter operations and keeps the intermediate out of HBM, while leaving the GEMMs unchanged. A hand-written Liger Triton kernel provides the same fusion without Dynamo guards or compile latency, trading Inductor’s shape-specialized speed for generic execution across changing shapes.
DiffusionGemma: The Developer Guide
DiffusionGemma is an experimental text-generation model and developer guide covering serving, customization, and non-autoregressive generation. Built on the Gemma 4 backbone, its 26B Mixture of Experts architecture activates 3.8B parameters during inference, refines 256-token canvases in parallel, and shifts the bottleneck from memory bandwidth to compute. Block autoregressive diffusion commits denoised 256-token blocks to the KV cache, combining bidirectional denoising with causal prefill for longer sequences; vLLM integration provides an OpenAI-compatible local server. A Sudoku fine-tuning example using Hackable Diffusion reports improvement from roughly 0% base-model success to 80%, while SFT reduces the demonstrated solve from 48 to 12 steps through adaptive stopping. The guide also identifies bidirectional attention and re-noising as mechanisms for global constraint handling and self-correction.
Migrating Your GitHub CI to Hugging Face Jobs
Trackio's GitHub Actions CI was limited by generic hosted machines, maintenance-related delays, and lack of readily available GPU runners, while the project needed CPU checks and CUDA-based tests. The setup keeps GitHub Actions as the CI coordinator and uses a dispatcher Space plus the huggingface/jobs-actions bridge to launch ephemeral GitHub runners inside Hugging Face Jobs, selected through labels such as hf-jobs-cpu-upgrade and hf-jobs-t4-small. The guide covers duplicating and configuring the dispatcher, creating a GitHub App, supplying runner credentials and an HF token, changing runs-on, choosing Docker images, and validating runs and logs; Playwright and NVIDIA CUDA images are used for CPU and GPU examples. In Trackio, the CPU job took 1m10s versus a 1m40s GitHub ubuntu-latest baseline, while the GPU check passed in 45s and cost less than a cent at the t4-small rate; logs were available through both systems.
Scaling up your microservice testing with Apache Spark—Part 2
Part 2 describes how to build a repeatable Apache Spark replay harness for testing high-impact microservices against historical, production-shaped inputs without reimplementing the service. The design turns dependency state—such as database snapshots, versioned rules, merchant configuration, eligibility state, and logged service responses—into datasets, while representing side effects as output rows and matching fidelity to the decision under test. A thin Spark wrapper reconstructs ReplayInput objects, invokes the same deterministic decision engine used online, compares current and candidate results, and writes traceable diffs with intermediate values and rule explanations. The source cautions that incomplete context, distributed state, asynchronous workflows, and external actors limit replay’s predictive power, so it should model explicit assumptions rather than claim full production simulation. It presents migrations as a practical point to establish these boundaries and use replay to reduce risk.
Top topics
- Architecture830
- Performance812
- LLMs648
- Machine Learning617
- Open Source561
- AI Agents542
- AI454
- Developer Experience446
Weekly digest
The best engineering reads, once a week.