Tech Meridian
← ENTITY INDEX
COMPANY · ENTITY #3140
Related event timeline, sources and context from the news index.
InfoQ published an article arguing that as AI moves from chatbots to autonomous agents the key engineering challenge is managing what an agent knows, remembers, and sees in real time. It recommends separating context, memory, and knowledge, building a 'context engine' or control plane to select and deliver minimal high-value context to LLMs, and optimizing context to reduce hallucinations and operational costs.
DoorDash reduced latency in its AI-powered grocery reorder flow from about 16 seconds to 2 seconds by precomputing recurring purchase candidates (from 90 days of history) with batch LLM inference, applying deterministic validation, and storing reusable shopping-list data; the serving path then resolves items against the live catalog. The change doubled shopping-list views and increased the flow’s order rate by roughly 34%, and it builds on DoorDash’s shared agent platform and grounded shopping interface (Ask DoorDash).
InfoQ's AI, ML and Data Engineering editorial team published a 2026 trends report and released a podcast episode where editors and guest Chip Huyen discuss key trends: a shift toward smaller language models and reasoning models, wider deployment of AI agents (from personal to organizational), interest in System One models such as TypeSafe AI's Jev for cheaper/faster inference, growing focus on AI evaluations, and advances in Physical AI. The panel also offered near-term predictions including more US robotics innovation, cost and energy constraints on infrastructure, broader use of specialized/mixed agents, and continued delineation of tasks best suited for AI versus humans.
At GOTO Copenhagen Charles Humble (interviewed by InfoQ) argued that AI sustainability should be a top-level architectural concern and urged practices such as using pre-built or smaller models, training on renewable energy, shifting demand to low‑carbon times/locations, and optimizing runtime and hardware. He cited IEA and cloud providers’ reports showing rising electricity use and emissions linked to AI and data centers, and suggested location- and time-aware strategies to materially reduce carbon intensity.
InfoQ announced three live online certification cohorts in November 2026: AI Engineering with Adi Polak (Nov 12, PST) and Lan Chu (Nov 13, CET), and a High-Performing Teams cohort with Olimpiu Pop (Nov 17). The programs cover production AI topics such as retrieval-augmented generation (RAG), context and memory pipelines, agent orchestration and control, AI infrastructure and evaluations, and team operating models; sessions are live, not recorded, and certification requires participation and a capstone.
An InfoQ article summarizes three paradigms for building agentic harnesses around LLMs — chains (linear workflows), graphs (explicit control-flow with nodes/edges), and code (sandboxed, execution-first systems) — and argues the field has moved from ad‑hoc glue to production-grade, observable architectures. The piece highlights tradeoffs (latency, fragility, resilience) and recommends treating LLM outputs as untrusted, decoupling planning from execution, and investing in observability and minimal necessary abstractions.
A Harvard study using Jellyfish data (≈300 million work events across ~700 firms, Jan 2021–Mar 2026) finds AI coding assistants modestly increase code output (lines +12%, commits +9%, PRs +5%; only commits statistically significant) while AI agents produce larger increases (lines +30%, commits +20%, PRs +23%). Despite more code, assistants did not increase resolved Jira issues and agents raised resolved issues by only 12%, with significant downstream impacts: average PR review time +49%, comments per PR +53%, share of PRs requiring changes doubled, and share of engineers performing reviews +14%—suggesting review and other complementary stages are bottlenecks.
At QCon AI, Manju Rajashekhar (founder & CEO of Mad Labs) presented “Actors for Agents,” arguing that the actor model maps naturally to the runtime needs of production AI agents. He described patterns — state ownership, message passing, write-through state, and supervision boundaries — to manage failures, continuations, and mutable execution graphs so agentic demos can become resilient production systems.
InfoQ released an expert-written article series titled "Agentic AI Architecture" that surveys emerging software-architecture patterns driven by LLMs and agentic systems, covering decision decomposition, failure modes, observability, context engineering, and implementation frameworks; the collection is available as a collated eMag PDF. Contributors include Mallika Rao, Karthik Ramgopal, Adi Polak, Subash Natarajan, Ahilan Ponnusamy, and Rafał Gancarz.
An InfoQ article argues that advances in LLMs and reasoning systems are driving a shift from microservices-based architectures to agentic systems that decompose decisions rather than functionality. Agentic systems introduce non-deterministic, context-driven execution paths, making decision quality a primary reliability concern and requiring patterns like bounded autonomy, evaluation loops, and deterministic fallbacks.
In an InfoQ podcast, Rachel Lee Nabors discusses local-first web apps and their Mima client (mima.social), arguing that on-device small language models (SLMs) such as Llama 3.2B can provide privacy-preserving, low-latency, and lower-cost alternatives to cloud-hosted LLM wrappers. She also covers evaluating models with golden datasets, using client-side AI for moderation and social aggregation, and prospects for OS-level personal agents and dynamic UI standards like A2UI.
An InfoQ webinar titled “AI in Production: What Breaks, What Works, and Who Approves It?” gathered ThoughtWorks practitioners and security/privacy experts to discuss operational challenges of agentic systems, including semantic and embedding drift, RAG leakage, model memorization, and the scaling gap for human code review as coding agents produce more production code. Panelists recommended treating evals as persistent infrastructure, shifting checks left in the harness, using mutation testing and continuous tracing, enforcing agent identities and tiered access, red‑teaming, and isolating secrets to limit blast radius.
InfoQ published an eMag, 'Evolutionary Architecture in Practice', featuring articles by participants of the InfoQ Certified Architect Program that present evolutionary architecture as a continuous feedback system and explore its intersection with AI — including a piece on agentic fitness functions (an AI judgment layer for CI/CD), system comprehension, platform guardrails, and federated ownership strategies.
Spotify reported monthly merged changes rose from about 8,100 to 17,000 year‑over‑year in August after expanding AI‑assisted development. The company added a background coding agent called Honk (running Claude via the Agent SDK in Kubernetes), expanded fleet automation and Fleetshift orchestration, and tightened monitoring, testing, rollout safeguards and rollback capacity after incidents where higher change volume strained verification controls.
In an InfoQ Architect's Podcast episode, Joy Ebertz (principal software engineer at Imprint) discusses how AI is changing software architecture and development practices — stressing the growing importance of unambiguous requirements and tests, the distinction between 'vibe coding' and tightly supervised pair-programming with AI for critical systems, and uncertainty about long-term patterns (e.g., microservices vs monoliths) and how to train junior engineers.
Researchers from Meta, MIT and the University of Washington propose “Context Language Models” (CLMs), an approach that treats model context as an editable file the model can rewrite to preserve facts, drop irrelevant information, and develop reusable context-management strategies. Evaluated on benchmarks (BrowseComp-Plus, EdgeBench, ContextBench) the paper reports substantial gains — e.g., zero-shot CLMs showed +11.4% accuracy with 21.5% fewer FLOPs on BrowseComp-Plus, and reinforcement learning improved Qwen3.5-9B on BrowseComp-Plus from 28.8% to 42.5% while reducing FLOPs — while also noting open risks such as irrecoverable deletions, prompt-injection channels, cache invalidation, and uncertain real-world robustness.
Google published Android Bench 2.0, a major update to its benchmark for evaluating AI models on Android development tasks that adds long‑horizon tasks (LHTs), agentic evaluation (using providers' agents), and a continuous, multi‑factor scoring system instead of binary pass/fail. The new dashboard lists recent models (e.g., Gemini 3.8/3.7, OpenAI GPT‑6, Anthropic Fable 5.1, Kimi K3, Qwen 3.8 Max); at the time of reporting Claude Opus 5.5 led LHT pass rates at 32%, with GPT 6 Astra at 28%.
A Coleman Parkes survey of 300 senior engineering leaders commissioned by Undo found that AI coding agents speed code generation but shift the main bottleneck to debugging, comprehension, and maintenance — respondents report averaging 9.8 hours/week producing code versus 16.9 hours/week debugging, 35% of AI-generated code reaches production before teams fully understand it, and 80% say agents struggle with complex problems in mission‑critical C/C++ codebases.
InfoQ is hosting a free 60-minute live webinar, “AI in Production: What Breaks, What Works, and Who Approves It?”, with five practitioners (Hien Luu, Katharine Jarmul, Zichuan Xiong, Premanand Chandrasekaran, Lan Chu) discussing production challenges for RAG, coding agents, verification, data exposure, and governance; registration is free and a recording is shared with registrants. The panel will explore agent autonomy, verification gates, sensitive-data handling, production RAG architectures, and how engineers make the internal case for controls.
An InfoQ article describes a Go-based design for preserving per-session ordering over Kafka using consistent hashing, a two-level worker hierarchy, and one goroutine per active session. The author explains in-place retries, contiguous watermark commits, and operational hardening (rebalance draining, backpressure, idempotency, dead-letter queues) and reports the system has processed >40M production messages with no observed ordering violations and achieved >100k send-side messages/sec in tests.
QCon London 2027 (April 13–16) will run 15 practitioner-focused tracks covering productionizing AI, architectural patterns, data platforms, observability, SDLC changes, and team/leadership issues, with over 75 speakers sharing real-world trade-offs and operational lessons. Tracks explicitly include evaluation and guardrails for agentic systems, security for human and agent failures, and integrating AI without sacrificing reliability or operability.
Akka tested a spec-driven workflow for AI-assisted software porting across 65 open-source projects, using a delivery harness that cycled through discovery, specification, porting, benchmarking, and improvement. The initial tranche took 99.3 hours and consumed 9.41 billion tokens; Akka reported lines-of-code or performance improvements in 57 of the 65 ports and used Claude with Akka Specify for implementation, with Sonnet averaging 61 minutes per port versus 120 minutes for Opus (Opus used ~40% fewer tokens).
InfoQ publishes a platform-engineering playbook based on a retail inventory recommendation system that reduced hallucination rate from ~15% to ~1.5% without changing the foundation model by applying platform-level controls (automated retry loops, intent-validation gates, prompt registries, tool authorization, and custom instrumentation). The article links a companion GitHub repo with runnable examples that use a local llama3.1 via Ollama connected through LiteLLM and describes routing through foundation models such as Gemini and GPT under enterprise traffic.
An InfoQ Live panel moderated by Renato Losio discussed how AI and automation are changing production operations and observability, turning operational data into actionable insights for incident response. Panelists from Genesys (formerly AWS), Netflix and Groundcover covered practical uses (e.g., ML/agents reducing MTTR), instrumentation like OpenTelemetry, and architectures such as eBPF and BYOC agentic observability.
InfoQ is running two five-week online certification cohorts in October 2026: AI-Assisted Engineering (starts Oct 19) focuses on building harnesses, permission limits, tests and CI checks for coding agents working on brownfield codebases; AI Security & Privacy Engineering (starts Oct 26) focuses on tracing sensitive data, threat modeling, red teaming and testing controls for AI products. Facilitators include Katharine Jarmul and Zichuan Xiong, and cohorts emphasize hands-on capstones, peer review, and comparing logged results with initial predictions.
At QCon AI, Forter principal engineer Ben Maraney described a two‑week internal sprint that gave ~200 R&D staff hands‑on experience building AI agents. The team lowered friction by running a custom MCP server, combining no‑code and code platforms, avoiding complex RAG setups, and coordinating with security and legal to speed adoption.
Docker announced Cloud Sandboxes: hosted, hardware‑enforced microVM execution environments that let developers run AI coding agents on Docker-managed infrastructure and move sandboxes between local machines and the cloud using the same CLI and isolation model as local Docker Sandboxes. Docker also updated the Sandbox Kit Specification to Kits v3, packaging kits as standard OCI images; commentators welcomed the flexibility for long-running and parallel agent workloads but warned about remaining attack surfaces when sandboxes access external services.
Mallika Rao, former engineering leader at Twitter, Walmart, and Netflix, presents at QCon AI about the operational and system-design challenges of adaptive recommendation systems. She emphasizes real-time feedback loops, retrieval freshness, multi-stage orchestration, and end-to-end latency and cost trade-offs required to continuously learn and evolve recommenders in production.
InfoQ hosted a podcast panel with Meryem Arik (DoubleWord), Clara Higuera Cabañes (BBVA) and Jeff Smith (C Proof) that surveyed enterprise AI adoption, covering inference/token cost reduction, tensions between proprietary models (e.g., Claude) and high-quality open-source alternatives, governance and ethics, the shift to non-deterministic "generative code," and speculative future architectures such as Kolmogorov-Arnold Networks.
An InfoQ article defines the 'agent harness' as the non‑model engineering around an AI model that makes an agent production‑ready, split into development (memory, tools, retrieval, orchestration) and operations (observability, guardrails, cost/control). It compares Harness‑as‑a‑Service (examples: AWS AgentCore) with self‑managed stacks (examples: LangChain + Agent Router/Envoy AI Gateway on Kubernetes), walks through building a sample agent (FinBot), and advises starting minimal and evolving the harness.
Allow Google Analytics to measure visits? It stays off until you choose. You can change your choice at any time.