Loading…
Latest reads
Every post, summarised
Grounded summaries of engineering blogs, newest first. Each links to the original.
- 4,671 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.
Latest posts
Together AI and Y Combinator partner to launch the first dedicated GPU cluster for the YC community
Together AI and Y Combinator are partnering to provide YC portfolio companies with a dedicated GPU cluster, addressing compute access as a growing bottleneck for AI-native startups. The service is designed for both inference and training, letting founders reserve, provision, and manage GPUs through Together’s self-service portal rather than making long-term commitments or routing usage through YC. Startups can run short-term sprints at long-term rates, use capacity from single-node workloads to larger deployments, and receive GPUs in minutes, while the cluster is already running at full utilization and companies can plan needs months ahead. Together says its research on attention mechanisms and the Mamba architecture targets faster inference and improved unit economics. The partners plan to expand the cluster and support startups as their requirements grow.
Run Ray on TPU, Part 1: The foundations
Ray 2.55 makes Google Cloud TPUs a first-class accelerator in Ray, with official APIs, pre-built images, and core-library support rather than the earlier experimental path. The central constraint is TPU slicing: a multi-host workload must land on one intact slice because its hosts communicate through the slice’s ICI (Inter-Chip Interconnect). On GKE, the Ray Operator installs KubeRay and a TPU webhook; the webhook labels hosts, while Ray Core uses those labels to reserve a whole slice atomically through the public slice_placement_group() API, with topology and accelerator version as inputs. The guide shows node selection by TPU generation and topology, a multi-host numOfHosts field, and a RayCluster manifest that brings up one worker per host. Ray AI libraries such as Data, Train, and Serve generally invoke this placement primitive themselves, while direct use targets custom distributed workloads; the API is public but alpha.
In-House LLM Serving at Netflix
Netflix established an in-house serving platform to run large language model inference directly inside existing production environments alongside traditional machine learning models. Built upon NVIDIA Triton Inference Server and vLLM, the unified architecture handles member-scale routing, candidate generation, feature fetching, and model execution over gRPC and OpenAI-compatible HTTP endpoints. The team shifted their primary engine from TensorRT-LLM to vLLM to support custom architectures, simpler debugging, and non-trivial constraint logic. Addressing production obstacles required patching Triton's frontend for guided decoding, pinning dependent library versions, rewriting logits processors in C++ for vLLM V1, and handling state machine resets during engine preemptions. The resulting platform unifies deployment flows while preserving operational stability across evolving model schemas.
Responsible AI: Governance, Principles, and Practical Guide
Responsible AI is presented as a lifecycle-wide practice for designing, developing, deploying, and monitoring AI systems with fairness, transparency, accountability, privacy, safety, and human oversight as requirements. The guide connects technical controls—secure encrypted data pipelines, documented dataset provenance, demographic bias audits, adversarial robustness tests, access controls, and continuous monitoring—with governance mechanisms including named model owners, cross-functional oversight, model-risk assessments, and immutable decision logs. For generative AI, it recommends output policies, training-data leakage testing, guardrails, and red-team testing, while model cards, automated fairness checks, independent audits, and incident response plans support transparency and accountability. Regulatory preparation includes mapping systems to the EU AI Act’s risk categories and documenting design, training data, and intended use; the NIST AI Risk Management Framework and OECD AI Principles are identified as governance references.
Tech builds on AI. Finance protects the margin.
AI-native tech companies must protect unit economics as agents accelerate changes in compute consumption, pricing, and revenue recognition, while gross margins remain below classic software levels. Finance teams built around extracts, spreadsheets, and monthly reconciliation can miss repricing changes, metering errors, and compute-commitment risk. The proposed foundation is an evolving ontology that keeps product, plan, usage, and cost meanings current, with Stripe data entering Unity Catalog through OpenSharing and Lakebase providing transactional Postgres on the lakehouse. Genie One uses that ontology to answer governed, sourced questions about gross margin, consumption revenue at risk, and compute spend, while people retain decision authority. The post describes organizations using Databricks to consolidate reporting, forecasting, workflows, and finance applications, positioning a shared data-and-AI platform as the path from an initial answer to an ongoing finance platform.
Building a soccer coaching app on Databricks
Coach’s Corner, also called La Pizarra, turns high-frequency soccer tracking data into a bench-side application for replay, tactical analysis, scouting, standings, and agent-generated dossiers. Built as a Databricks App, it ingests NDJSON feeds at 25 frames per second through Auto Loader and Spark Declarative Pipelines, enforcing 46 data-quality expectations across bronze, silver, and gold layers. Liquid clustering supports 1–3-second DBSQL queries, while Lakebase synchronizes gold data to Postgres for millisecond replay reads and separates sequential playback from exploratory analytics. The scouting layer grounds Genie, Vector Search, a Unity Catalog-registered xG model, and an Agent Bricks supervisor in governed data, with Claude calls routed through the Unity AI Gateway, MLflow tracing, and a deterministic fallback. Together, these components are presented as a way to deliver traceable insights within seconds without forcing coaches to interpret raw tables or analysts to relay every result.
The cost of saying yes has changed
Generating initial code patches has become significantly cheaper with AI agents, shifting the primary expense of small feature requests from writing code to debating scope in meetings. Teams can use agent-generated patches as diagnostic probes rather than final deliverables, turning abstract scope debates into concrete artifacts that reveal true system touchpoints and risks. However, low generation costs do not translate to low ownership costs, as changes touching areas like authorization, compliance, or public contracts still demand significant human review. Constrained attempts allow engineers to price uncertainty quickly and shift scope discipline from pre-implementation speculation to evidence-based code review.
Eclipse Dataspace Components on AWS: Cost optimization strategies
Deploying Eclipse Dataspace Components (EDC) connectors on AWS presents challenges in forecasting infrastructure expenses without clear benchmarks. Estimated costs differ substantially between business-critical setups and non-critical environments based on compute and database sizing. A business-critical configuration relying on Amazon Aurora PostgreSQL db.r6g.large and standard Amazon ECS with AWS Fargate totals an estimated 387.00 USD monthly in the eu-south-2 region. Sizing down to db.t4g.medium and switching to AWS Fargate Spot lowers the non-critical monthly expense to 164.00 USD, representing a 58% reduction. Core cost drivers remain baseline infrastructure components, while storage, API calls, and data transfers contribute marginally under baseline assumptions.
Meta’s Spark Muse 1.1 is now available on Databricks, fully governed by Unity AI Gateway
Databricks announces support for Meta’s Muse Spark 1.1 through Model Provider Services (MPS) in Unity AI Gateway, addressing fragmented API keys, access controls, and usage visibility when organizations adopt new models. An MPS is a Unity Catalog securable that stores provider configuration and an encrypted API key, while callers use their Databricks credentials and the gateway attaches the key at request time. The post demonstrates registering Muse Spark through the OpenAI provider type with Meta’s API base URL and Responses API, then governing use with Unity Catalog privileges, model allowlists, policies, rate limits, usage metering, and inference tables. Requests are routed through the gateway, where access and guardrails are applied before reaching Meta; usage, spend, tokens, latency, status codes, and optionally full payloads are recorded for attribution and audit.
Decoding Customer Value: pLTV Models for India’s Quick Commerce Delivery Wars — part 1
Swiggy’s pLTV effort addresses early-stage lifetime-value prediction for Food Delivery and Instamart, where sparse signals, zero-inflated outcomes, long tails, and trial-driven acquisition make conversion models unsuitable. It encodes 350+ pre-order signals across 12 groups in a simple multilayer perceptron with three shared hidden layers, task-specific heads, a custom MAE loss, and Adam optimization. A four-head variant added order-count predictions for each vertical, producing marginally better performance with 135K rather than 363K parameters. Evaluation emphasized ranking through Spearman correlation and decile diagonal coverage, reaching above 0.75 Spearman and 70–80% diagonal coverage. In production bidding experiments, the in-house model delivered better retention and higher gross order value per acquired user at no incremental platform cost, while a third-party platform acquired more users per dollar.
DE Dost: Teaching Machines to Be a Friend on the Last Mile
DE Dost is a GenAI onboarding companion embedded in Swiggy’s delivery-executive app, designed for new DEs who receive information but lack confidence-building, immediate guidance. It supports questions in nine languages across SOPs, earnings, payouts, app navigation, shifts, and support, adapting detail and tone to conversational context. Before the first LLM call, a graph state assembles user context, preprocessing resolves references, and embedding-based semantic routing separates general knowledge-base questions from personal diagnostics. Retrieval uses authored QnA pairs plus hybrid semantic and lexical scoring to handle paraphrases and terms such as “MG” and “COD,” while guardrails constrain follow-ups to available knowledge. Evaluation spans about 800 cases per language across fleet and payout combinations, with macro accuracy trending at 90%+ and a target P50 latency of 4.5 seconds after reducing an initial 15-second pilot.
How We Made All Swiggy Apps Operations AI-Native
Swiggy’s mobile engineering teams faced operational overload as releases, production signals, and reviews required switching among Jira, Confluence, Bitrise, Play Console, Crashlytics, New Relic, Google Sheets, ImageKit, and Slack. The response was an operating layer built around Slack-native agent workflows, reusable AI skills, MCPs, and scheduled jobs connected to release, observability, QA, cost, and internal mobile systems. It can generate readiness updates, group Jira work, identify missing owners and blockers, summarize vitals, analyze alerts, follow up TRM actions, and recommend RAMP, HOLD, or ROLLBACK for staged rollouts by correlating Crashlytics, Play Vitals, and adoption. Humans still approve rollout actions, while recurring channel reports make observability and accountability more consistent and reduce clerical integration work. The stated model lets agents handle mechanical load while engineers retain judgement, ownership, product quality, architecture, and execution.
Enable Your SAML Requesting App for Cross App Access
Cross App Access allows enterprise applications federated with Security Assertion Markup Language to interact with third-party resource APIs without migrating to OpenID Connect. The flow adapts the Identity Assertion Authorization Grant specification to SAML by adding three token exchange steps following the standard single sign-on handshake. First, the application exchanges the validated SAML assertion for an OAuth refresh token using Okta's token endpoint and a private key JSON Web Token. The application then uses the refresh token to request a short-lived Identity Assertion Authorization Grant token from Okta. Finally, the application presents the Identity Assertion Authorization Grant token to the third-party authorization server to obtain an OAuth access token.
Improvements to Cursor in Slack
Cursor in Slack now gives users more control over tasks by presenting a plan before work begins and reporting status as it progresses. Its Slack presentation has also changed: compact footer links replace in-message buttons, while tables, pull requests, and artifacts render more cleanly. In named multi-repo environments, Cursor can select an environment with access to related frontend, backend, and shared-code repositories; if it needs another repository, it prompts for a switch and resumes the task afterward. Cross-channel workflows let Cursor read context from other channels and threads and send updates to the original thread or a relevant channel. Together, these changes support earlier redirection, broader repository access, and task coordination across Slack workspace conversations.
The New DMARC is Here
DMARCbis is finalized as three Standards Track RFCs: RFC 9989 for the core protocol, RFC 9990 for aggregate reporting, and RFC 9991 for failure reporting. Existing DMARC records continue working as before, so most senders do not need immediate changes. The largest behavioral update replaces Public Suffix List lookups with an algorithmic DNS Tree Walk, in which receivers query _dmarc records up the DNS hierarchy; the psd tag can declare hierarchy boundaries. The pct tag is replaced by boolean t, while np can specify a policy for non-existent subdomains, including np=reject. Reporting knobs rf and ri are removed, but rua and ruf remain available; senders using fractional pct should use a none-to-quarantine-to-reject rollout, keep reading reports, and avoid rushing to reject.
Q&A: How Capcom Brought Path Tracing to RE ENGINE Across PRAGMATA and Resident Evil Requiem
Capcom integrated path tracing into RE ENGINE across PRAGMATA and Resident Evil Requiem to deliver unified indirect and direct lighting. Over two years, the team built a reference path tracer validated against standard DCC tools, scaling it down into a real-time version optimized for NVIDIA DLSS Ray Reconstruction denoising. To manage rendering performance, Capcom avoided expensive AlphaTest texture lookups by creating a ScreenSpaceAlphaTest method comparing ray-hit positions against screen-space depth. Sampling noise in emissive and indoor regions is mitigated using Next Event Estimation and a custom implementation of ReSTIR GI. Replacing traditional shadow maps with continuous path tracing reduced the visual gap between gameplay and cinematics while enabling real-time light transmission through strand hair.
Your AI is ready. Your data foundation probably isn’t
Cushman & Wakefield’s enterprise AI program addresses fragmented experiments, disconnected data, and uneven organizational maturity across a 53,000-person workforce. Over four years, Chief Digital and Information Officer Sal Companieh used a product operating model, business-linked accountability, co-created investment decisions, and shared architecture standards to build a common foundation while preserving business-unit flexibility. Databricks supports that strategy as a partner and platform, with its intelligence layer and Genie helping employees query trusted data in natural language, examine quality and governance, and monitor compliance. The company says the time from idea to outcome has fallen from months to days, while client and acquisition onboarding has materially accelerated. Companieh identifies human behavior, education, and trust—not technology alone—as essential to making change durable.
From experiment to insight: how Dotmatics Luma and Databricks make AI-ready science a reality
Scientific workflows generate data across instruments, teams, and partner networks, but siloed systems can strip away metadata, lineage, and experimental context. Luma, Dotmatics’ scientific intelligence platform, captures outputs continuously and harmonizes them into structured, FAIR-compliant records, while Databricks supplies scalable storage, governance, and enterprise data and AI infrastructure. Running natively on Databricks, Luma preserves a continuous digital thread across experiment design, acquisition, analysis, reporting, and sharing, with Delta Sharing supporting governed exchange with collaborators. Chromatography illustrates the approach: connectivity across vendor systems and Virscidian’s Analytical Studio automate LC/MS processing while retaining context and adding dashboards, registration, and management tools. In one pharmaceutical deployment, Luma connected about 1,500 of more than 5,000 instruments across four LC/MS vendors, enabling cross-vendor performance trending, unified purity analysis, and a foundation for AI and machine learning.
How To Set Up a Staging Environment on Heroku in 3 Easy Steps
The guide explains how to add a Heroku staging environment so changes are tested in a near-identical, isolated replica before production. It uses a pipeline with production and staging apps, disables production automatic deployments, connects staging to GitHub, and enables automatic builds, with Review Apps offered for pull requests. The setup applies Twelve-Factor practices by separating config vars and databases, recommending migrations into a fresh Postgres database or warning that pg:copy overwrites staging data and can create compliance risks. After validation, slug promotion sends the exact compiled staging asset to production without recompiling, while numbered releases support rollback; database changes must still be reversed separately.
The skills gap behind agentic AI — and how Databricks is closing it with a new context engineer certification and agent trainings
Databricks introduced the Databricks Certified Context Engineer Associate beta exam to validate skills for building reliable, production-grade AI agent systems as organizations scale agentic AI. Context engineering is presented as the practice of curating, maintaining, and filtering tokens, memory banks, and tool parameters so an LLM can solve a task; the certification is intended to benchmark that technical skill set. Databricks also added AI Agent Fundamentals, Building Retrieval Agents on Databricks, and Agent Evaluation, covering agent reasoning, retrieval-augmented architectures, and systematic performance testing and improvement. Its AI-first certification-prep guide is available on every certification page, works with free tiers of major LLMs, includes guardrails and disclaimers about hallucinations, and shows candidates how to use Free Edition for hands-on practice; registration for the Context Engineer Associate is open, with the first exam scheduled for July 29, 2026.
Top topics
- Architecture834
- Performance816
- LLMs660
- Machine Learning622
- Open Source563
- AI Agents553
- AI470
- Developer Experience448
Weekly digest
The best engineering reads, once a week.