Customer Engineering Manager

Building a more trustworthy Cortex for enterprise software delivery. Enterprise teams do not need AI that merely writes code. They need AI that fits into real software-delivery practices: clear ownership, dependable validation, controlled access, useful continuity, and evidence that helps people decide what should happen next. This Cortex update advances that goal. It strengthens the end-to-end coding-agent experience - from request through code change, validation, review, recovery, and learning - while keeping developers and organizations in control. From code generation to accountable delivery. Cortex is evolving from a coding assistant into a more complete development partner. It can help teams understand requests, prepare code changes, run appropriate checks, summarize outcomes, and present the resulting work for review. The experience now makes the status of a coding task easier to understand at a glance. Developers can see the files involved, the scope of a change, validation status, and whether a change is still awaiting review. This is important in enterprise environments, where a natural-language statement that work is "done" is not enough. Teams need a concise, reviewable record of what the agent actually produced and what evidence supports it. Clear separation of responsibilities. Reliable AI development requires different capabilities to have clearly defined roles. Cortex now more deliberately separates the work of producing code, carrying out bounded development actions, validating results, and providing advisory review. This helps reduce confusion between an agent proposing a change and a system confirming that the change is ready for use. The benefit is a more predictable workflow. Coding assistance can remain focused on solving the task; validation can provide concrete execution evidence; and review can supply an independent perspective. Each signal is visible to the developer rather than being hidden behind an opaque automated decision. Human judgment remains the control point. Cortex now includes an advisory Verifier capability for eligible users. After a proposed change has been prepared and validation evidence is available, the verifier can provide an additional assessment: ready to proceed, needs revision, reject, or requires further review. This is designed as an advisory control, not an autonomous authority. Developers and teams retain the ability to inspect the diff, review supporting evidence, rerun validation, accept or reject a proposal, request a different approach, or escalate an important decision. Cortex does not silently merge code, authorize a patch, or substitute an automated verdict for human responsibility. That model gives organizations a practical path to use AI more broadly without weakening code-review standards or governance expectations. More honest validation and risk signals. A passing validation command is valuable evidence, but it is not a complete statement of quality. A patch can pass a local test while still needing review for coverage, maintainability, security, compatibility, or business correctness. Cortex now makes this distinction clearer. It separates validation outcomes from review recommendations and highlights when certain local checks were unavailable. Teams can see both what was verified and what remains uncertain. This creates a more useful decision-making experience than a simple pass/fail label. It helps developers move quickly when evidence is strong while preserving appropriate caution when the evidence is incomplete. Faster, more resilient AI workflows. AI-assisted development must be efficient as well as capable. Cortex has strengthened routing and recovery behavior so coding tasks are handled by the most appropriate capability and do not consume excessive time or resources through unnecessary retries or unsuitable fallback paths. The result is a more dependable experience for common coding requests and clearer behavior when a supporting service is unavailable. Instead of leaving a developer waiting indefinitely or presenting confusing results, Cortex keeps recovery bounded and communicates the practical outcome. When an optional advisory review service is unavailable, the code change remains available for user review. Internal provider details are not exposed in the developer experience, and no automated approval is implied. Reliable continuity across the development experience. Development work often spans restarts, refreshes, reconnects, devices, and long-running sessions. Cortex has improved session continuity so active work is less likely to be lost, duplicated, or displayed inconsistently when a client reconnects. Conversation restoration is now more reliable across desktop, browser, and mobile experiences. The improvements focus on preserving meaningful user and agent activity, maintaining clearer timing and review context, and preventing incomplete or operational-only records from appearing as customer-facing conversation. Cortex has also expanded support for longer active conversations. This better reflects real enterprise development, where an investigation, implementation, testing cycle, and review may extend well beyond a short chat exchange. Consistent review across Cortex surfaces. Whether a developer is working from an editor, browser, or mobile device, the underlying expectation is the same: proposed changes should be understandable and reviewable. Cortex is aligning the change-review experience across supported surfaces so users receive clearer status signals, compact validation indicators, accessible review actions, and consistent handling of pending changes. This reduces friction when work moves between environments and helps teams maintain a common review standard. Enterprise access that follows the user. Enterprise AI must respect the organization's existing boundaries. Cortex is strengthening the connection between identity, access, and AI capability so that developers receive the assistance appropriate to their role, plan, and approved work environment. This gives organizations greater confidence that advanced coding assistance, connected enterprise systems, and higher-impact workflows are available deliberately, not simply because a client requests them. It also creates a cleaner path for security, IT, and engineering leaders to expand adoption in line with internal policy. Controlled access for advanced capabilities. Advanced AI review capabilities need to be introduced with clear access controls. Cortex now supports subscription-based availability for advisory patch review, allowing organizations to make these capabilities available according to their commercial plan and governance requirements. This enables a measured adoption path. Organizations can begin with standard coding assistance, enable enhanced review for the right teams, and expand usage as internal practices and confidence mature. A connected development experience. Modern development does not happen in one place. Engineers move between their IDE, browser, and mobile device as they investigate issues, review work, coordinate with colleagues, and return to active tasks. Cortex is making that experience more continuous. Developers can carry meaningful work context across supported surfaces, return to previous discussions more easily, and maintain a more consistent experience regardless of where they choose to work. The goal is to reduce the friction of switching environments without weakening governance or creating disconnected AI experiences. A foundation for continuous quality improvement. Cortex is also establishing a stronger evaluation foundation for coding-agent quality. This enables the platform to assess workflow reliability over time: how coding tasks recover, how often validation is performed, how review signals align with outcomes, and where user experience can improve. These capabilities support future improvements to coding agents, validation workflows, and verifier models while keeping data governance central. Organizations can control whether eligible feedback contributes to learning and evaluation, preserving transparency and choice. What this means for enterprises. Cortex is becoming a more accountable AI development environment. It helps teams accelerate coding work while retaining the safeguards that matter in production software delivery: * Clear evidence of what changed and what was checked * Better separation between generation, execution, validation, and review * Advisory AI review with human-controlled decisions * More honest visibility into validation status and remaining risk * Faster, bounded recovery when supporting services fail * Reliable continuity across editor, browser, and mobile workflows * Subscription-aware access to advanced capabilities * A scalable foundation for measuring and improving AI quality The objective is not automation for its own sake. It is to help engineering organizations deliver software faster with stronger visibility, clearer accountability, and greater confidence in every AI-assisted change.

The new playbook for engineering operations: insights from Cortex. Session summary. In this session, Ganesh Datta, CTO & Co-founder of Cortex, introduced the DRIVE Framework, a practical operating model for evaluating engineering organizations in the age of AI. Rather than focusing on developer productivity alone, Ganesh challenged leaders to think about operational health as a competitive advantage - sharing concrete examples of how leading engineering organizations are redesigning reviews, governance, and quality systems to keep pace as AI dramatically accelerates software development. Key takeaways. * AI has changed the bottleneck - engineering leadership hasn't caught up yet. As AI dramatically increases the speed of software development, the limiting factor is no longer writing code - it's ensuring organizations can maintain reliability, quality, security, and operational discipline. Engineering leaders need new operating models built for an AI-native world. * Developer productivity is no longer the metric that matters most. AI makes it easier than ever to produce more code, but shipping more software doesn't necessarily create more customer value. Leaders should increasingly measure operational health, customer outcomes, and system reliability instead of individual engineering output. * Operational Excellence Reviews should become a regular leadership practice. Ganesh introduced the DRIVE Framework as a structured way for engineering organizations to periodically assess operational health, identify systemic risks, and prioritize improvements before problems become outages or organizational bottlenecks. * Every engineering organization should intentionally design its own operating system. The highest-performing teams don't blindly copy Google, Netflix, or Amazon. Instead, they define the standards, governance, workflows, and engineering practices that best fit their own business, customers, and stage of growth. * AI increases the need for engineering standards - not less. As AI-generated code becomes commonplace, organizations need stronger ownership models, review processes, architectural standards, and service governance to prevent operational complexity from growing faster than engineering teams can manage it. * Focus on the metrics your customers actually experience. Rather than optimizing internal engineering dashboards alone, Ganesh encouraged leaders to evaluate customer-facing outcomes such as reliability, service health, and functional availability - the measures that ultimately determine whether engineering is delivering value. * Automate routine work so engineers can focus on high-leverage decisions. One example discussed automatically merging low-risk pull requests because senior engineers had become overwhelmed reviewing AI-generated code. As development accelerates, organizations should redesign workflows so human expertise is reserved for the highest-risk decisions. * Operational health is a systems problem - not an individual performance problem. The DRIVE Framework encourages leaders to evaluate how well the entire engineering organization functions together. Improving isolated teams or individuals has limited impact if the broader delivery system remains constrained. * The best engineering reviews create learning - not blame. Operational reviews should surface patterns, encourage cross-functional discussion, and drive continuous improvement rather than becoming compliance exercises or performance audits. The goal is to strengthen the organization, not assign fault. * The organizations that win in the AI era will pair AI acceleration with operational discipline. Ganesh's central message was that AI is becoming table stakes. The lasting competitive advantage will belong to companies that combine AI-powered development with strong engineering systems, clear governance, and a culture of continuous operational improvement.