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
Intelligent security at ClickHouse speed: How Cogent Security built an AI-native vulnerability management platform
Enterprise security defenders face shrinking exploitation windows alongside massive volumes of vulnerability findings across disconnected tools. To overcome query bottlenecks from traditional relational databases, Cogent Security designed a lakehouse serving architecture using Apache Iceberg and per-customer ClickHouse instances. The architecture relies on query-driven denormalized tables, custom compression codecs, and pre-computed projections to serve hundreds of millions of rows with sub-second latency. On top of this data layer, Cogent implemented an agentic text-to-SQL system that iteratively queries live data, references an in-house Ontology Service for semantic context, and applies majority-voted evaluations. This combination improved text-to-SQL accuracy from 40% to 94% while providing sub-three-second response times across 98.7% of queries at 100 million rows.
A New Framework for Evaluating Voice Agents (EVA)
EVA is an end-to-end framework for evaluating conversational voice agents across task accuracy and spoken conversational experience. It simulates multi-turn live-audio conversations with a user simulator, the evaluated voice agent, deterministic tool execution, validation, and metrics derived from recordings, transcripts, and tool-call logs. The framework reports EVA-A for Accuracy and EVA-X for Experience, combines code-based and LLM- or LALM-as-Judge metrics, and uses pass@3 and pass^3 across three trials to capture peak performance and consistency. Its initial release contains 50 reproducible English-language airline scenarios and 15 tools covering rebooking, itinerary changes, cancellations, standby, and vouchers. Results for 20 cascade and audio-native systems show a consistent tradeoff: agents stronger at task completion tend to provide worse experiences, while a substantial gap between pass@3 and pass^3 indicates inconsistent success.
Smarter Auto-Scaling for ClickHouse: The Two-Window Approach
ClickHouse Cloud required a vertical auto-scaling mechanism that preserves rapid scale-ups while avoiding prolonged over-provisioning after traffic drops. Operating on a single 30-hour lookback window forced scale-downs to take up to 30 hours, whereas an earlier fixed-factor CPU algorithm caused severe oscillations. The engineering team deployed a dual-window recommender pairing a 3-hour window for responsive scale-downs with a 30-hour window for single-step scale-ups. Alongside a trend-based hunting check to resolve conflicting window recommendations, the team introduced target-tracking CPU scaling using geometric mean watermarks and 10-minute median smoothing. This dual-window architecture reduced scale-down latency to 3 hours, prevented scaling oscillations, and significantly reduced infrastructure costs for variable database workloads.
Building high-performance full-text search for object storage
ClickHouse redesigned its native full-text index to address latency bottlenecks caused by random read patterns on remote object storage. Instead of using a Finite State Transducer dictionary, the new architecture organizes indexed tokens into fixed-size sorted blocks compressed with front-coding. Each data part separates index data across a small in-memory sparse index file, a sequential dictionary file, and a posting list file. This layout allows the query engine to pinpoint token locations with minimal random I/O and resolve multiple full-text functions directly from index files. Consequently, ClickHouse Cloud maintains predictable search latency across shared object storage while matching local disk performance and scaling through parallel query execution.
Fast regex search: indexing text for agent tools
Agentic coding has revived grep as a critical way to gather source-code context, but ripgrep can stall on large monorepos because it scans every file. The proposed remedy is a local inverted index of overlapping character trigrams, with regular expressions parsed into required and alternative trigram constraints that narrow candidate files before exact matching. The index stores posting lists in one disk file and a sorted hash-to-offset lookup table in another, memory-mapping only the table for binary-search lookups; hash collisions can broaden results but cannot produce incorrect ones. Freshness is prioritized because agents often search for text they have just written, and the client index layers user and agent changes over a Git commit. The reported conclusion is that instant local search creates qualitative workflow improvements, especially for bug investigation and large Enterprise repositories.
Structured Logging in .NET with Serilog and ClickHouse
Traditional logging solutions like Seq, ELK, and cloud platforms often face scaling bottlenecks, including slower queries, higher operational costs, and complex pipeline management. To address this, an ASP.NET service can write structured logs directly to ClickHouse using Serilog and the Serilog.Sinks.ClickHouse library. The sink transmits data over HTTP and provides a fluent C# builder to configure table schemas, typed JSON columns, full-text indexes, and MergeTree engine parameters. ClickHouse stores log data in a columnar format that delivers 10 to 20 times compression while enabling direct SQL queries, JSON property filtering, and real-time aggregations through materialized views. Furthermore, the sink includes an in-memory buffer to handle transient network drops with configurable batching, retry mechanisms, and callback hooks for failure alerting.
Build a Domain-Specific Embedding Model in Under a Day
The tutorial presents a pipeline for adapting a general-purpose embedding model when it misses fine-grained distinctions in domain documents and labeled query–document data is unavailable. It generates synthetic question–answer pairs with NeMo Data Designer, filters them for quality, mines hard negatives with a 95% positive-score margin ceiling, and unrolls multi-hop queries for contrastive training. The recipe fine-tunes Llama-Nemotron-Embed-1B-v2, evaluates base and adapted models with BEIR metrics, and exports the result to ONNX or TensorRT. On the reported datasets, it improved Recall@10 and NDCG@10 by more than 10%; Atlassian increased Recall@60 from 0.751 to 0.951 on JIRA data. The complete workflow can run in roughly 2–3 hours for about 500 documents, using a single 80GB Ampere-or-newer GPU.
Making Ads Count: Using MMoE and Auxiliary Tasks to Better Connect Buyers & Sellers
Etsy’s Ads Search ranking model needed a stronger purchase-intent signal because clicks are noisy and purchases are sparse later in the buyer journey. It retained its shared feature-representation and interaction layers, then replaced the shared feed-forward network with Multigate Mixture of Experts (MMoE), using task-specific gates and towers for click-through rate (CTR) and post-click conversion rate (PCCVR). Add-to-cart was added as an auxiliary task, while favorites were rejected after producing negative transfer. Offline, the resulting model increased Purchase and Click PR AUCs by average 3.5% and 1%, respectively, and Purchase ROC AUC by 0.5%. Online, it improved purchases and purchase calibration, supplied better inputs to auto-bidding, kept model size flat through pruning, and reduced inference cost.
Transformation in action: Raising the bar for customer experience
AI is shifting support from reactive backlog management toward proactive experience design, as teams gain capacity through faster, more efficient service across languages and time zones. The report says 58% rank experience and satisfaction as their top 2026 priority, up from 28%, while mature deployment triples the share able to always meet expectations from 9%. Intercom’s support team expanded access, improved first-response rates, and added onboarding sessions while absorbing a 300% demand increase without expanding headcount, enabled by deep integration of Fin. Clay uses Fin across Slack, email, and in-app chat, where it handles half of roughly 7,000 monthly queries and participates in 90% of all queries. The recommended model treats support as a product, combines AI with human handoffs, anticipates needs, builds trust, and personalizes interactions.
Transformation in action: Why ROI becomes clearer with deeper integration
Customer service teams often see early AI gains in response speed and efficiency, but deeper operational integration produces larger returns and makes value easier to measure. Research cited in the 2026 Customer Service Transformation Report says 62% of support teams have improved metrics after implementing AI, rising to 87% among teams at mature deployment; the share able to measure ROI rises from 35% to 70%. As deployment matures, teams shift from measuring time and cost savings to tracking how freed capacity supports higher-value customer work and revenue-generating activity. The proposed AI-first economic model treats AI as a workforce capability that improves human productivity, system improvement, revenue influence, and organizational agility. Intercom’s Fin illustrates the model: it is involved in 97% of customer conversations, resolves 83.5% end-to-end, and helped absorb more than 300% demand growth without scaling headcount.
Transformation in action: Why AI deployment depth makes the difference
AI adoption is widespread in customer service, but the report identifies a deployment gap between organizations experimenting with the technology and those integrating it deeply into operations. Based on a survey of 2,470 global support professionals, it finds that 82% of senior leaders invested in AI in the past year and 87% plan to invest in 2026, while only 10% report mature deployment. Mature teams put AI at the core of workflows, give it meaningful responsibility, and use continuous improvement to achieve higher-quality, more consistent support; 43% report that outcome, nearly double the rate for teams at initial deployment. Lightspeed illustrates the approach by combining tailored training, executive alignment, content investment, and AI operations discipline, reaching 72% Fin resolution without human intervention across 43,000-plus monthly requests and 12-plus languages.
Querying DateTimes in ClickHouse
ClickHouse provides specialized functions to filter, bucket, and compute durations on native DateTime columns without requiring external processing tools. Using the New York City taxi dataset imported into a MergeTree table, queries bucket pickup timestamps into hour and 15-minute intervals using toStartOfHour and toStartOfFifteenMinutes while casting values to the Time type to group across dates. The dateDiff function calculates journey durations in minutes, which enables trip speed calculations across time windows and powers inline ASCII visualizations generated by the bar function. By combining toDayOfWeek with conditional aggregations like countIf and the lag window function, queries uncover distinct morning surge trajectories between weekdays and weekends. The analysis demonstrates that weekday trip volumes peak around 8:45 AM alongside dropping speeds, whereas weekend volumes increase steadily across the morning.
Migrating Etsy’s database sharding to Vitess
Etsy migrated its long-running MySQL shard management to Vitess vindexes after an architecture spanning about 1,000 tables and shards, 425 TB of data, and roughly 1.7 million requests per second. The legacy ORM used a single index database to map records to shards, creating a scaling bottleneck and single point of failure. To avoid moving existing data, engineers built custom SQLite lookup and hybrid vindexes, changed new shard assignments to match Vitess’s hash algorithm, and introduced routing incrementally table by table with traffic ramps, rollback, and query compatibility testing. After five years, about 2,500 pull requests, and 6,000 queries, the migration removed the index database dependency, made scaling operations take days rather than months, simplified developer access, and kept database performance largely unchanged without a mass data migration.
How Slack Rebuilt Notifications 📣
Slack redesigned its legacy notification architecture to address user confusion, conflicting desktop and mobile preference systems, and tight coupling between notification content and push delivery. The engineering team unified four separate preference models into a single framework with three channel options: All new posts, Mentions, and Mute. To ensure safe rollback and backwards compatibility without direct database mutations, Slack implemented a read-time translation strategy alongside a new desktop_push_enabled preference and auto-saving cross-platform React UI components. This decoupling of activity from push delivery resulted in a fivefold sustained increase in settings engagement, reduced reliance on per-channel overrides, and eliminated synchronization mismatches across mobile and desktop clients.
Kaizen for the AI era: How small improvements build smarter support
The post presents Kaizen as a structured approach to continuous improvement, centered on small changes, immediate iteration, and participation from people closest to the work. At Fin, this philosophy is applied to Agent Fin through the Fin Flywheel: teams train it on complex queries, test simulated conversations, deploy it across channels, and analyze performance with AI-powered insights. Each interaction feeds a repeatable loop rather than a one-time launch, while Human Support representatives examine escalations, ask why Fin could not resolve them, and identify content, process, or tooling fixes. A lightweight AI-powered logging process captures suggestions in the moment, enabling subject matter experts or frontline teams to implement changes quickly; together, the two systems make continuous learning an ingrained capability and are presented as a source of competitive advantage.
Supabase Joins the Stripe Projects Developer Preview
Supabase has joined the Stripe Projects developer preview, a Stripe CLI workflow for provisioning real services and obtaining working credentials with a single command. The integration provisions a complete Supabase project—including Postgres, Auth, Storage, Edge Functions, and Realtime—without requiring dashboard interaction, while supporting both developers and AI agents through deterministic, repeatable steps. After installing the Stripe CLI and adding supabase/project, stripe projects env --sync writes credentials to a local .env file and stores them in the Stripe Project for agent access; existing Supabase users can link their account. Projects remain in the user’s Supabase account with dashboard, connection-string, and data access, and database credentials can be rotated with stripe projects rotate supabase/project. The feature is a developer preview, and Stripe and Supabase are requesting feedback through Discord or GitHub.
From firefighting to building: How AI agents restored our team’s core productivity
Grab’s Analytics Data Warehouse team supports more than 1,000 monthly users and 15,000 tables, but repetitive questions about definitions, lineage, quality, and basic enhancements consumed about 40% of engineering time. It deployed a multi-agent AI system that routes Slack requests into enhancement or investigation pathways, using FastAPI and LangGraph for stateful, cyclical collaboration, Redis for caching and sessions, and PostgreSQL for persistent memory. Specialized agents query Trino, Hive, or Delta Lake, inspect GitLab code, check incidents and pipeline health, or prepare semi-automated code changes, while human reviewers approve production modifications. The system autonomously resolves most standard inquiries and some common enhancements, reducing resolution time by an order of magnitude and bringing simple answers to users within minutes. The resulting capacity shift reclaimed several FTEs worth of engineering bandwidth and moved hundreds of hours from reactive support to roadmap work.
Introducing Composer 2
Cursor has made Composer 2 available as a coding model in Cursor. It is priced at $0.50/M input and $2.50/M output tokens, with a faster variant priced at $1.50/M input and $7.50/M output tokens. The release reports higher scores across CursorBench, Terminal-Bench 2.0, and SWE-bench Multilingual, with Composer 2 scoring 61.3, 61.7, and 73.7, respectively. Its training combines a first continued pretraining run with reinforcement learning on long-horizon coding tasks, allowing it to handle challenges requiring hundreds of actions. Composer 2's standard pricing and included usage pool on individual plans are presented alongside the faster option, which is set as the default.
Building a Magic Mirror: AI retail experiences with Remix
Shopify’s “magic mirror” is an AI-powered retail installation that uses a display and hidden webcam to deliver personalized messages, animations, and product interactions in under a minute. A Remix route handles the browser and server boundary: a button-triggered POST captures an image, sends it with a curated prompt to a vision-capable language model, and renders a compliment plus a makeup recommendation; server-sent events optionally let a remote device trigger or reset the experience. The prompt combines persona and safety constraints, a plain-text product and inventory table, and a required JSON schema, while strict product-ID lookup provides graceful degradation. The interface uses explicit states, delayed capture, timed response and reset screens, and animated transitions. The production setup keeps brand configuration, shared components, actions, and SSE infrastructure reusable across mirror modes.
March 2026 newsletter
The March 2026 ClickHouse newsletter details major platform updates, community deployments, and architectural practices for high-throughput analytics. Recent product announcements include ClickHouse 26.2 featuring time-based block flushing for streaming feeds, general availability of native inverted indexes for full-text search, and the release of chDB 4 with a Pandas-compatible DataStore API. Community implementations show Cloudflare managing 1.61 quadrillion daily events, while Trigger.dev isolated multi-tenant SQL execution using a domain-specific query language that compiles safe ClickHouse queries. In data optimization techniques, Hookdeck accelerated webhook payload search from over 30 seconds to under 400 milliseconds by partitioning payload hashes into typed bucket columns and scanning iterative time windows. Additional operational guidance emphasizes setting the ttl_only_drop_parts parameter and aligning table partitioning with time units so background merges drop entire data parts cleanly without triggering mutation rewrites.
Top topics
- Architecture834
- Performance816
- LLMs660
- Machine Learning622
- Open Source563
- AI Agents553
- AI470
- Developer Experience448
Weekly digest
The best engineering reads, once a week.