Detailed Report of Agentic Engineering and Harness engineering at Zalando , Uber , Netflix, AirBnB, Anthropic and others
This report covers 51 publicly documented agentic and harness engineering initiatives current to 2 September 2026: 39 at the five focal companies (Zalando, Uber, Netflix, Airbnb, Anthropic) and 12 more from Block, Google, GitHub, Stripe, Microsoft, LangChain, METR and research groups — each with lifecycle scope, harness detail, metric where one was published, and source tier. Counts are documented public examples, not totals of internal systems, and are not comparable across companies.
Agentic engineering is the organizational and software-delivery discipline of structuring work so agents can plan, act with tools, inspect feedback, and iterate toward an outcome. Harness engineering is the surrounding execution and control layer that turns model capability into dependable work: environment, tool interface, context/memory, orchestration, observability, verification/evaluation, and governance/security. Not every initiative below is fully autonomous — the record spans pipeline automation (Zalando's migration tool), interactive copilots (Gemini CLI, Claude Code), background agents (GitHub's coding agent, Headless Goose), and managed software factories (Uber, Stripe Minions). They are included because they show the evolutionary path toward agentic systems.
Zalando's public record shows a layered internal platform: a LiteLLM API proxy serving 2k monthly active users on just six small pods, chat and CLI surfaces, and production pipelines on top. A risk-based bot auto-approves 33% of PRs classed as low-risk; an LLM-powered UI component library migration hit about 90% accuracy at under $40 per repository; and the SRE postmortem pipeline delivered a 3x productivity boost in summaries and processed annual analysis in under 24 hours. An LLM-as-a-judge framework assures search quality across languages before market launches.
Uber's primary engineering posts document DragonCrawl blocking ten high-priority bugs in three months while saving thousands of developer hours, and PerfInsights' dual validation (with LLMCheck) cutting false positives from over 80% to the low teens, plus a 33.5% drop in validated static-analysis findings in four months. Newer scale figures — 5,000+ unit tests/month from Autocover, about 1,800 code changes a week and 95% engineer reach for Minion, 7x weekly active user and 9.4x request growth for the Software Factory — come from secondary reporting and should be treated as claims, not verified primary figures.
Netflix emphasizes specialist paved paths. Its performance-agent workflow reads production profiles, checks out the exact commit running in production, matches against a markdown pattern catalog, runs unit tests and canary deployments, and sends a human-approved code review — with CPU opportunities of 0.5–4.6% found across seven services. Metaflow's spin lets a coding agent build and test ML workflow steps incrementally. The oci-agent case, whose estimate reached only 25% of a polished baseline, shows why specialist paved paths and actor-critic checking matter.
Airbnb's strongest primary case is the Enzyme-to-React-Testing-Library migration: work broken into per-file parallel steps with retries, richer context, and feedback loops migrated 75% of target files in four hours during one stage and completed nearly 3.5K files in six weeks — versus 1.5 engineer-years estimated manually. Newer AirDev isolated remote environments, AirChat agentic coding, async agent execution, and CI-for-agents details come from job listings and secondary reporting.
Anthropic publishes the most reusable harness playbook: tool/ACI design, context management, and long-running work split between an initializer agent and coding agents that leave durable artifacts between sessions. Sandboxing for Claude Code — OS-level filesystem and network isolation — reduced permission prompts by 84% while shrinking the attack surface, a rare public metric for governance rather than throughput.
Synthesizing the 51 initiatives, one architecture recurs. Each layer is grounded in documented practice.
Classify work before an agent touches it. Zalando's risk bot routes 33% of PRs to auto-approval; GitHub scopes its coding agent to low-to-medium complexity tasks in well-tested codebases.
Repository instructions, retrieval, ownership and dependency data. Netflix's performance agent checks out the exact production commit and reads a markdown pattern catalog; Metaflow uses CLAUDE.md instructions; Airbnb's migration succeeded by adding richer context on retries.
Bounded tasks, retries, budgets, durable state. Airbnb decomposed 3.5K files into per-file parallel steps; Anthropic separates an initializer agent from coding agents with artifacts between sessions; the Manage–Execute–Audit loop keeps task state outside the growing context.
Clear semantics, minimal surface. Stripe exposes a curated subset of 400+ MCP tools per agent because over-tooling degrades performance; Anthropic's guidance centres on agent-computer interface design.
Ephemeral sandbox or worktree, scoped filesystem and network. Anthropic's OS-level isolation cut permission prompts 84%; GitHub's agent boots a VM per task with limited internet; Microsoft's THINKINGBOX isolates MCP tool sessions; Airbnb's AirDev provides remote isolated environments.
Tests, static analysis, telemetry, canaries, independent evaluators, human approval. PerfInsights' dual LLM validation cut false positives from 80%+ to the low teens; Netflix runs unit tests plus canary CPU comparison plus human review; LangChain's self-verification loop and METR's advisor-actor-rater scaffold show the actor-critic pattern.
Traces, cost and latency, audit, secrets, policy, kill switches, continuous evaluation. Uber's uReview benchmark tracks precision, recall, cost and latency per review; Google's Agent Quality Flywheel uses OTel traces, AutoRaters and failure clustering; Zalando's proxy centralizes access and cost.
migrate legacy codebases, refactor and simplify complex logic, generate unit tests, streamline dependency upgrades, and speed up triage workflows
Running in an automated background environment when an issue or PR is created and tagged
A manager to maintain the task state and determine the next subtask, a fresh-context executor to perform it, and a read-only auditor to verify the resulting environment state before the next round.
Boots a VM, clones the repo, pushes commits to a draft pull request, and tags developers for review. Limited internet access to trusted destinations, CI/CD workflows require human approval before running, and existing branch protections apply.
orchestration layer, multi-step reasoning chains, external tools such as linters or test runners, context retrieval, and error recovery
GenAI evaluation service, User Simulator, AutoRaters, custom rubrics, Automatic Loss Analysis, OTel traces
Async generator loop, ModelRouterService with pluggable strategies, Scheduler state machine for tool execution, and hybrid loop detection via SHA-256 hashing and adaptive LLM self-checks.
generating dynamic, incident-specific single pane of glass interfaces to reduce manual data gathering during incidents
PreCompletionChecklistMiddleware
1. Prompt the model to analyze the present situation and generate advice. 2. Have the model generate 6 possible actions it could take next (3 of which are generated with the advice in context, and 3 of which are not). 3. Tell the model to rate each of the 6 po
isolated MCP-compatible tool sessions, complete execution traces, and outcome evaluation over terminal backend state.
deterministic harness nodes for CI and linting, agentic nodes for implementation, a harness-enforced limit of two CI rounds before escalation to a human, and a curated subset of over 400 MCP tools exposed per agent
| Company | Initiative | Lifecycle scope | Public metric | Tier | Source |
|---|---|---|---|---|---|
| Zalando | Automated UI component library migration | Migration and Refactoring | less than $40 for each code repository; accuracy of about 90% | Primary | engineering.zalando.com |
| Zalando | Chat UI and CLI | coding | — | Primary | engineering.zalando.com |
| Zalando | CLI tool | maintenance scripts for model access in the terminal and coding tasks | — | Primary | engineering.zalando.com |
| Zalando | intelligent SRE assistant postmortem analysis pipeline | Operations | 3x productivity boost in summaries | Primary | engineering.zalando.com |
| Zalando | LiteLLM API proxy | coding | 2k MAU with just six small (2k CPU, 4 GB) pods | Primary | engineering.zalando.com |
| Zalando | LLM-as-a-judge search quality assurance framework | Evaluating and assuring search quality at scale with multi-language support for pre-market launches and existing markets | — | Primary | engineering.zalando.com |
| Zalando | Multi-stage LLM pipeline | Operations | Processed annual data analysis in under 24 hours | Primary | engineering.zalando.com |
| Zalando | Python based migration tool | Migrating partner facing applications from one of the UI component libraries to the other one | — | Primary | engineering.zalando.com |
| Zalando | Risk-based PR approval tool | code review | 33% of our PRs are low-risk and are auto-approved by the bot | Primary | engineering.zalando.com |
| Uber | Autocover | generating unit tests | generating 5,000+ unit tests per month | Secondary | newsletter.pragmaticengineer.com |
| Uber | DragonCrawl | Testing | blocked ten high-priority bugs from impacting customers while saving thousands of developer hours and reducing test maintenance costs | Primary | uber.com |
| Uber | LLMCheck | Testing and Validation | Achieved a 33.5% reduction in validated static analysis findings in four months (from 265 down to 176). | Primary | uber.com |
| Uber | Minion | ships about 1,800 code changes a week | about 1,800 code changes a week, and 95% of Uber's engineers use it | Secondary | sandeep.dev |
| Uber | PerfInsights | Testing and Optimization | Reduced false positives from over 80% to the low teens. | Primary | uber.com |
| Uber | Software Factory agentic offerings | code review, self-healing CI failures, completing E2E PRs with visual validation, triaging on-call alerts, debugging inc | weekly active users grew 7x, weekly agentic requests grew 9.4x | Secondary | x.com |
| Uber | uReview | AI code review for all pull requests | Handles code review for all pull requests | Secondary | x.com |
| Netflix | AI agent-assisted performance engineering workflow | Automated performance profiling, bottleneck identification, pattern/anti-pattern recognition, and code review generation | CPU and latency savings observed in production; found across seven different services with savings between 0.5 to 4.6% of CPU cycles. | Secondary | youtu.be |
| Netflix | Metaflow spin | Incremental development, testing, and debugging of ML and AI workflow steps using AI coding agents | — | Primary | netflixtechblog.com |
| Netflix | oci-agent | Analysis and Evaluation | 25% of the baseline estimate | Primary | netflixtechblog.com |
| Netflix | Spin | development and debugging | — | Primary | netflixtechblog.com |
| Airbnb | @generateMock | GraphQL data mocking at scale | — | Primary | medium.com |
| Airbnb | AirChat | agentic coding and software development tasks | — | Secondary | markolukicic.substack.com |
| Airbnb | AirDev | providing isolated remote development environments for concurrent agentic work | — | Secondary | markolukicic.substack.com |
| Airbnb | async agent execution platform | build, validate, review, and ship high-quality software | — | Secondary | careers.airbnb.com |
| Airbnb | CI platform infrastructure | create, build, test, review, and merge PRs for AI coding agents | — | Secondary | careers.airbnb.com |
| Airbnb | LLM-driven code migration | Testing | updated nearly 3.5K React component test files from Enzyme to use React Testing Library (RTL) instead in just 6 weeks | Primary | medium.com |
| Airbnb | LLM-driven migration pipeline | updating nearly 3.5K React component test files from Enzyme to use React Testing Library (RTL) instead | migrated 75% of our target files in just four hours | Primary | medium.com |
| Anthropic | Claude Agent SDK | coding and environment setup | — | Primary | anthropic.com |
| Anthropic | Claude Code | coding | — | Primary | claude.com |
| Anthropic | Claude Code auto mode | Automated coding tasks and tool execution while delegating permission approvals to model-based classifiers | — | Primary | anthropic.com |
| Anthropic | Claude Code on the web | coding and execution verification | — | Primary | anthropic.com |
| Anthropic | Claude Cowork | Collaborative developer workflow and productivity | — | Primary | anthropic.com |
| Anthropic | Code Review | Code review for pull requests | — | Primary | anthropic.com |
| Anthropic | coding agent | to resolve SWE-bench tasks, which involve edits to many files based on a task description | — | Primary | anthropic.com |
| Anthropic | Coding Agent for SWE-bench | coding and testing | — | Primary | anthropic.com |
| Anthropic | Computer use reference implementation | execution | — | Primary | anthropic.com |
| Anthropic | Filesystem isolation | coding and execution verification | — | Primary | anthropic.com |
| Anthropic | Network isolation | coding and execution verification | — | Primary | anthropic.com |
| Anthropic | Sandboxing for Claude Code | coding and execution verification | 84% reduction in permission prompts | Primary | anthropic.com |
Method: 51 publicly documented agentic/harness engineering initiatives compiled to 2 September 2026 — 39 at Zalando, Uber, Netflix, Airbnb and Anthropic, 12 from other organizations — each with lifecycle scope, harness detail, public metric where disclosed, source tier and URL. Counts are documented public examples, not internal system totals; metrics use each company's own definitions and windows and are not cross-comparable. Some initiative descriptions were shortened for space; lifecycle categories in the matrix were assigned from each row's stated scope.