Loading…
Latest reads
Every post, summarised
Grounded summaries of engineering blogs, newest first. Each links to the original.
- 4,668 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
Palana (Part 1): Why Grab built a secure platform for autonomous AI Agents
Palana is Grab’s Kubernetes-native platform for running autonomous and semi-autonomous agents with isolation, controlled access, and operational oversight. The post describes why laptop-based tools and simply running agents in containers do not adequately address persistent state, identity, credentials, network access, auditability, and emergency control. Its architecture gives each agent a namespace, service account, storage, network policy, and Vault scope, while Envoy, ext-authz, and OPA mediate and log HTTP(S) egress; proxy-only secrets keep underlying credentials out of agent processes. A portal and pcli manage lifecycle, and external kill switches and idle shutdown provide controls outside the agent. Palana is used for development environments, Slack automation, framework testing, and long-running internal workloads, with deeper architecture details reserved for Part 2.
Beyond the warehouse: How METRO Markets built a do-it-all data platform on ClickHouse Cloud
METRO Markets outgrew its self-hosted Impala and Kudu-based Hadoop stack as the marketplace expanded, with parallel stores, inconsistent reporting, and too much maintenance burdening a lean data team. They evaluated BigQuery, Snowflake, and ClickHouse against production queries, finding faster ClickHouse performance, including one response falling from four seconds to under 200 milliseconds, while its pricing supported more predictable costs. Airbyte moves CDC data from MySQL and Postgres into ClickHouse; ReplacingMergeTree handles deduplication during background merges, and BigQuery-hosted Google Analytics data enters through cloud storage and external tables. The resulting warehouse consolidates company data and supports seller analytics, credit risk features through Query API endpoints, observability, and emerging AI use cases, while teams query a shared dataset through compatible tools; planned Kafka and Flink integration would extend it to streaming workloads.
Data scientists: Powering the future of AI and analytics
Data scientists connect statistics, programming, and domain knowledge to turn raw data into models, recommendations, and decisions with business consequences. The role now extends beyond classical modeling to large language models, generative AI applications, agentic systems, and production workflows involving deployment, monitoring, and retraining. Modern practice combines Python, SQL, and libraries such as scikit-learn, PyTorch, Spark, and MLflow with data engineering basics, communication, and domain expertise. Data scientists contribute across problem framing, data access, preparation, feature engineering, experimentation, deployment, and lifecycle monitoring, with governed access and lineage helping align training data with production data. The article concludes that AI assistants and agents can automate routine work, but human judgment remains necessary for trustworthy evaluation, business framing, and actionable recommendations.
Why everyone is talking about real-time analytics (a note from “the yellow company”)
Real-time analytics is presented as a foundational workload for customer-facing analytics, observability, fraud detection, operational applications, and AI agents rather than a niche use case. The discussion argues that performance claims require open, transparent, reproducible benchmarks that disclose methodology, datasets, queries, results, and operating conditions, including continuous ingestion, concurrency, memory pressure, and fresh data. A real-time database should support queryable data within milliseconds to a couple of seconds, incremental transformations, bounded tail latency as writes increase, configurable read-time versus write-time work, unified support for text, vector, JSON, and structured analytics, and efficiency as workload scales. The article connects renewed attention to agentic workloads, whose continuous, high-volume querying makes low latency, concurrency, and fresh context increasingly necessary, and calls for shared standards beyond a single latency number.
MosaicLeaks: Can your research agent keep a secret?
MosaicLeaks examines privacy leakage in deep-research agents that combine private enterprise documents with external web retrieval. Queries that appear harmless individually can reveal sensitive facts when an observer combines them over multiple hops, producing intent, answer, or full-information leakage. The benchmark contains 1,001 multi-hop chains interleaving local and web questions, and compares baseline behavior with task-only reinforcement learning and Privacy-Aware Deep Research (PA-DR). Task-only training raises strict chain success from 48.7% to 59.3% but increases answer/full-information leakage from 34.0% to 51.7%, whereas PA-DR reaches 58.7% success with 9.9% leakage by rewarding stage-specific decisions and penalizing query-level privacy risk. The evaluation is controlled, using synthetic enterprise documents, a fixed web corpus, three company contexts, and one multi-hop question-answering harness.
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%.
Appcues delivers personalized customer engagement with ClickHouse Cloud
Appcues replaced an analytics stack that included Snowflake and Airflow after rising scale pushed queries to 20 seconds or more and limited reports to 10-minute windows. The company evaluated ClickHouse Cloud on AWS for real-time segmentation, exporting Snowflake data to Parquet in S3 and testing complex email segments with nested conditions, historical events, and negative logic. Production events now pass through Kafka and a clickwriter consumer, while a clickclient library built on Elixir’s Ecto generates controlled ClickHouse queries; SharedReplacingMergeTree, table engines, and materialized views were tuned during a one-year migration. Across 1.31 PB and 410 billion events, P95 queries fell below two seconds, ingestion latency dropped to about five seconds, and analytics costs decreased 23%, while new funnel reports and AI-powered insights were added.
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.
What's new in Postgres Managed by ClickHouse: RBAC, Terraform, ClickPipes, extensions, and more
Postgres managed by ClickHouse, which entered public beta in May, received a broad update covering console operations, per-service RBAC, Terraform, Stripe provisioning, ClickPipes migrations, billing, capacity reservations, and extensions. The redesigned console adds service health, resource, and query views, while RBAC provides View and Manage roles across the console and public REST API; the alpha Terraform provider covers lifecycle operations including resizing, HA, replicas, restore, and password rotation. Postgres-to-Postgres ClickPipes now automate schema migration, usage is integrated into ClickHouse Cloud’s metering and billing, and AWS capacity reservations are allocated dynamically across availability zones. Underneath, pg_clickhouse and pg_stat_ch moved to the header-only clickhouse-c library, removing a C++ exception conflict, using Postgres memory contexts, and reducing the x86-64 Linux extension binary from 4.9 MB to 1.4 MB; the update shipped less than a month after beta, with Terraform and extensions still evolving.
How Stagwell built privacy-safe ID matching on Databricks
Stagwell built a Databricks Marketplace App to let brands match first-party records against its Identity Spine without sending raw customer data outside their own workspaces. It combines Databricks Clean Rooms, Unity Catalog, Jobs and Notebooks, Delta Sharing, and a React/Express interface. Workspace OAuth and On-Behalf-Of authorization enforce the user's Unity Catalog permissions, while separate service principals and a scoped personal access token support app and clean-room operations. Within the clean room, a packaged notebook joins brand data with the Identity Spine, resolves multiple identifiers, and computes match rates, coverage metrics, and household and consumer IDs; results include aggregated audience insights for activation, while raw customer data and proprietary matching code remain separated.
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.
What is Artificial Intelligence (AI)?
Artificial intelligence (AI) is a branch of computer science that enables machines to perform tasks associated with human intelligence, including learning, reasoning, pattern recognition and decision-making. Modern AI generally learns patterns from large datasets, tunes internal weights and parameters during training, evaluates outputs on held-out data, and applies the resulting model during inference to classify, predict, generate content or trigger actions. Most organizations fine-tune existing foundation models rather than train from scratch, while output quality remains dependent on the completeness, bias and quality of training data. The page separates reactive machines and limited memory from theoretical theory of mind and self-aware systems, and distinguishes today’s narrow AI from theoretical general AI and superintelligence. It also describes generative AI, common applications, risks including hallucinations, bias, privacy and security gaps, and governance, concluding that practical adoption depends on real problems, trusted data and responsible oversight.
Data Engineering for AI: A Practical Guide for Data Professionals
Data engineering for AI extends traditional ETL by adding requirements for model-ready data, unstructured inputs, feature engineering, governance, and production monitoring. The guide addresses data professionals building or scaling AI infrastructure, covering ingestion, architecture, data quality, privacy compliance, generative AI, and career development. It emphasizes shared feature schemas, pipeline data contracts, lineage, statistical validation, drift detection, and PII stripping or anonymization as responsibilities shared across data engineering and data science. For generative AI systems, it describes preparing RAG pipelines by ingesting and chunking documents, creating vector embeddings, and indexing them for semantic retrieval, while evaluating vector databases for latency, scale, and integration. It concludes that reliable AI depends on fresh, accurate, compliant data and ongoing operational and architectural review.
Data Warehouse Types: A Complete Guide to Architectures and Use Cases
A data warehouse is a centralized repository for structured data, supporting complex queries, reporting, and business intelligence rather than transaction processing. The guide compares architectures by scale, latency, cost, scope, ownership, and governance. Enterprise data warehouses integrate organization-wide sources through ETL, apply cleansing and validation, and provide a governed source of truth, while data marts focus on departmental analysis and may be dependent or independent. Operational Data Stores replicate current or recent operational data for reporting refreshed from minutes to hours, whereas virtual, cloud, hybrid, and lakehouse designs trade physical consolidation, scalability, flexibility, and governance differently. The comparison also frames lakehouses as combining open-format data lake flexibility with warehouse-style governance and transactional reliability.
The new bottleneck
AI coding tools have significantly lowered the cost of generating software, yet many engineering organizations fail to realize overall delivery speed improvements. Applying the Theory of Constraints reveals that eliminating code production as a bottleneck shifts inventory directly into surrounding legacy processes that remain unadjusted. New friction points consistently emerge across underspecified requirements, prolonged design handoff gates, senior engineer review capacity, and external sign-offs from legal or security. Organizations can address these blockers by interrogating legacy agile ceremonies checkpoint by checkpoint to determine if their original constraints still exist. Practical remedies include adopting real-time co-development between product and engineering, treating initial designs as fluid starting points, and restructuring processes around running rapid, high-volume experiments.
What is Row-Level Security?
Row-level security (RLS) is a database access control that filters which table rows a user may read or change according to identity, role, or session context. It supports multi-tenant SaaS, regional segregation, departmental access, compliance, healthcare, and partner use cases, while typically layering table-level grants with column-level security or data masking. The database applies a policy or predicate automatically at query time across BI dashboards, notebooks, ad-hoc SQL, APIs, and third-party tools, with separate USING and WITH CHECK rules for reads and writes. On Databricks, Unity Catalog row filters use SQL user-defined functions and apply across Databricks SQL, notebooks, jobs, and connected BI tools. The source recommends simple predicates, indexed keying columns, identity-based testing, and explicit enforcement for both reads and writes.
How A2A is Building a World of Collaborative Agents
A2A, launched one year before this June 18, 2026 update, is presented as a common language for conversational, dynamic agents that need to collaborate and hand off tasks securely rather than behave like stateless tools. The protocol uses black-box handoffs to keep sensitive data and proprietary processes encapsulated, prevent context pollution, support clarifying exchanges, and distribute specialized workloads across teams or services. Foldrun demonstrates this model by exposing protein-structure prediction as a standalone agentic interface that manages long-running tasks, adjusts parameters based on prediction confidence, and selects among AlphaFold 2, OpenFold 3, and Boltz-2 depending on the molecule. In an A2A-compliant environment, users can register Foldrun and delegate work in natural language without custom glue code, while a primary agent handles other research tasks. The post also cites applications in commerce, streaming, IT operations, and regulated telecom, while Python and Go SDKs are 1.0 GA.
Beyond LoRA: Can you beat the most popular fine-tuning technique?
Parameter-efficient fine-tuning (PEFT) reduces the memory required to adapt open models, enabling fine-tuning with a fraction of the memory and, in some cases, quantized models. LoRA is dominant because it freezes base weights and trains a small set of added parameters, but its popularity may also reflect ecosystem momentum and uneven comparisons. The post presents shared-condition benchmarks for language-model math reasoning and image generation, using identical models, datasets, code, hardware, and measures including test performance, VRAM, forgetting or drift, runtime, and checkpoint size. The results show that LoRA performs well but is not consistently best: other techniques can outperform it on one or more axes, especially in the image-generation benchmark. The recommended choice is therefore to compare techniques against a use case’s needs and constraints, including quantization support, layer coverage, adapter merging, and downstream serving compatibility.
Is it agentic enough? Benchmarking open models on your own tooling
Final-answer checks can show whether an agent completed a task, but they miss the turns, tokens, seconds, errors, and command path required. The proposed tool-specific benchmark measures that work on transformers, comparing models, revisions, tasks, and three access tiers: bare installation, source clone, and a packaged Skill containing CLI documentation and examples. It uses deterministic tasks with exact matching, launches one Hugging Face Job per model–revision–task combination on identical hardware, and records traces in a Hugging Face Bucket. Reports include match rate, median new, cached, and generated tokens, time, error rate, and marker adoption, while traces expose the agent's commands. On transformers, the CLI plus Skill helped large open models but caused Qwen3-14B to misinterpret documentation as an executable tool, dropping classify-sentiment match from 100% to 0% in the cited comparison.
June 2026 newsletter
ClickHouse’s June 2026 newsletter marks the project’s ten years as open source, noting more than 2,000 contributors and community projects alongside engineering updates. It profiles Vasily Chekalkin’s 26.5 improvements to WebAssembly user-defined functions, including the DETERMINISTIC keyword and visibility in system.functions, and covers native Jaeger storage, the ClickCannon benchmark tool, and Open House announcements. The 26.5 release reports Top-N optimizations that make ORDER BY … LIMIT through JOIN 20.4× faster with 175× less memory on TPC-H SF100, while GROUP BY … LIMIT is 11.9× faster with 185× less peak memory. Other featured systems include a 15-million-rows-per-hour KPI pipeline, join-engineering work yielding a 26× improvement on join-heavy TPC-H workloads, and community guidance on inserts, columnar storage, ArrowStream, and RAG ingestion.
Top topics
- Architecture834
- Performance815
- LLMs659
- Machine Learning621
- Open Source563
- AI Agents553
- AI469
- Developer Experience448
Weekly digest
The best engineering reads, once a week.