Peter Nelson
Jan 23, 2026 - May 12, 2026
42
Decisions
0
Active Todos
8
Patterns
Decisions (42)
Open strategic review of RLC/RLK identity + upstream binding (Dirty Frag triggered)
Saturday 5/9 in #department-heads, in immediate response to Justin's Dirty Frag status table and Nathan's note about CIQ patches being shared with the RESF, Peter announced he wants the leadership team to take up a strategic question next week: what recurring vulnerabilities imply about CIQ's kernel posture, how tightly to bind to upstream, how to work with the RESF, and what it means going forward to be RLC and RLK. Aimed at framing input for the mid-to-late June LA in-person product-strategy session with Bjorn and Greg.
Commit engineering to vuln-handling infra/automation at Leadership Roundtable
At the 5/11 Leadership Roundtable, Peter accepted an explicit action item to prioritize vulnerability-response infrastructure and automation work in engineering, and to update Chris Baek as the interim process owner. The commitment converts the 5/8 internal-to-engineering commitment (build/test infra to eliminate reactive interrupts) into a cross-functional commitment with Bjorn, Greg, Chris, and Lindsay in the room.
Coach Mariah toward sharper, peer-level feedback posture — Reno trip offered
Following a heated DM disagreement about how to handle a senior IC's return-from-leave conversation, Peter offered Mariah an explicit choice: be treated the way Peter treats Bjorn and Greg (sharp, direct, unvarnished disagreement) or stay softer. Peter framed the conversation as investment, named that Mariah is undersized for where she could be ('I would like to see you pulled more into the core of things here'), and offered to be in Reno end of next week (Thursday) to discuss over lunch.
Ryan to POC AI-driven Veeam image builder, gated on Justin-approved test suite
Ryan will POC an AI-driven image builder, starting with Veeam, that automates custom image builds, testing, and documentation. Hard gate: AI-generated images must pass a Justin-approved test suite to prevent hallucinations. Long-term vision: a 'Chipotle line' image configurator letting users compose custom, repeatable builds from validated components.
Tighten Jira-as-system-of-record into active enforcement — instruct teams to ignore Slack-only requests
All significant work must be in Jira to count as a commitment. Peter will instruct teams to actively ignore requests that exist only in Slack. This escalates the Apr 18 quality decision from policy ('ticket your work') to enforcement ('we will refuse to act on un-ticketed requests'). Justin owns the enforcement in Build/Test/Deployment, the function most contaminated by ad-hoc Slack asks.
Ryan Smith owns docs.ciq.com
docs.ciq.com was orphaned after Gwen's departure. In Ryan's 1:1, Peter assigned ownership to Ryan, who will coordinate with Arian and Steven on the existing work.
Sensitive Decision
Ship CIQ kernel patch with extra fix; contribute upstream; race to be first/best on CVE response
Linux kernel CVE response: CIQ shipping 10 fixes vs CentOS Stream's 9 (CIQ found and is fixing an extra issue related to the CVE). Extra commit submitted upstream to centos-stream and acknowledged for inclusion. CIQ pushing to be first EL distro to release, with primary goal of customer reassurance and secondary goal of public proof point that CIQ contributes to security and is large enough to serve big customers. Also pushing patches to RLC kernels as fallback in case RH doesn't move quickly.
Defer ARM64 Pro Hardened build until Core42 commits — group decision Peter endorsed
In Apr 26 Sovereign AI response review meetings, the team — with Peter participating — decided the response language to Core42 will acknowledge that Pro Hardened on ARM64 (and FIPS-143 ARM certification) is contingent on a client commitment, not unilateral CIQ investment. ARM64 build estimated weeks not months once committed; FIPS-143 ARM is ~$200k / 4-6 months and gates on a deal commitment. Peter explicitly told the room: "We are going to need Nathan to say when. I am not going to be able to say on this call."
Sensitive Decision
Block Max's calendar — don't schedule anything for the time being
In a DM (4/17 11:10 AM Pacific), Peter directed Sarah (assistant) to stop scheduling anything for Max Spevack for the time being. This operationalizes the indefinite-leave posture triggered by Christina's 4/16 7:45 PM call telling Peter that Max needs time away. Rather than letting Max's calendar keep generating missed-meeting signals to the org (like the 4/16 1:1 and interview no-shows), Peter chose to block new scheduling entirely until further notice.
Manage Wesley via Greg; set boat-anchor trigger with Chris (preserve optionality)
In the Chris W 1:1, Peter decided to manage Wesley's performance situation through his Greg relationship rather than moving Wesley onto a PIP or active weekly-1:1 performance track like Cole and Thomas Chin. Assessment: Wesley is capable but slow, requires handholding, won't operate at senior/principal level, but is not a net negative. Chris asked to proactively flag if Wesley becomes a boat anchor on team productivity — that is the threshold for escalation.
Open hiring for a dedicated Ascender engineer
In the Design Sync, Peter took the action item to draft an Ascender Engineer job requisition and start the hiring process. Root cause: engineers Jimmy and Larry rejected a UI update PR for Ascender Pro citing the internal Quantic design system is not open source — an objection that is irrelevant since Ascender Pro is a closed-source commercial product. Bjorn will handle the immediate Jimmy conversation next week. Peter's move is structural: create an engineering owner whose role explicitly covers the closed-source Ascender Pro commercial mandate, so the category of blocker goes away.
Quality initiatives must be ticketed visible work, prioritized by Product
In the Brian/Brady weekly sync, Peter reinforced that quality cannot live as implicit expectations — Product must define and prioritize quality initiatives as explicit tickets that compete for resources against feature work. Engineering will only prioritize what is tracked. Companion frame: Exit Criteria are the product promise (Product owns, Engineering can challenge via debate); QA is delivery validation, split between Engineering (general releases) and Ryan's org (customer-specific fixes in mirrored environments). Brady to split test automation from build automation into a high-priority CI/CD ticket. Peter to verify with Justin that the build process at minimum runs a boot test.
Core42: pivot from Fuzzball sale to full-stack compliance partnership
After the Core42 Tech Dive Part 2 surfaced Core42 wants a single OS vendor for their full UAE compliance stack (NIST 800-53, BIS, IDAM, physical security) across three EOY-2026 GPU clusters, Peter immediately convened an internal Impromptu Zoom to reposition the opportunity. CIQ will propose a comprehensive partnership framing CIQ as the only group that can provide all requirements, with RLC Pro Hardened + Fuzzball + Ascender as the core stack and partners filling the remaining ~20%. Consultative play: CIQ will advise Core42 on which requirements in Eric Grundstrom's doc would cause unacceptable performance degradation vs. which can be met. Nathan to draft the proposal doc by EOD Saturday so CIQ can deliver an answer by Monday.
Shaped Ryan's H-1 strategic plan with specific direction
Reviewed Ryan's H-1 plan and gave specific strategic direction: reframe team skill tracking to outcome metrics, loop Product into GTM translation validation, create simplified 4-page sales pitch deck, align with Max on Rocky Linux upstream strategy (pure RESF vs hybrid best-of-breed) before engaging community, and present AI-assisted goal-setting process at Engineering Weekly Sync.
Expanded NVIDIA licensing ask to include Fabric Manager + NVISM
Expanded the NVIDIA licensing negotiation scope beyond DOCA-OFED to include Fabric Manager, NVISM, and other critical InfiniBand components not currently in the CUDA bundle. Sent email to Scott Hara while Bjorn was drafting the DOCA-OFED amendment language.
Committed to aggressive NVIDIA GPU Operator self-certification timeline for GTC
Led GPU Operator Self Certification meeting with NVIDIA. Committed CIQ to building pre-compiled GPU driver containers for Rocky Linux (mirroring Ubuntu model) and pursuing self-certification targeting preliminary completion by end of next week to support a GTC announcement.
Commit to benchmark accuracy sign-off by tomorrow
Committed to signing off on the accuracy of the RLC-AI benchmark document by Feb 21, gating its release. Greg conditioned approval on numbers being '100% accurate' and verifiable. Bjorn shared both internal and external draft documents for review.
Approve Jason Rodriguez performance exit path
Approved Nathan's plan to tell Jason Rodriguez he's 'not meeting expectations' and create an exit path. A 1-month exit plan is acceptable if Jason agrees. Performance gaps include missed deadlines (shim review for Rocky 9 late, Rocky 7 review pending) and unshippable code (large PR unlikely to pass review).
Core42 Technical Assessment - Multi-Product Positioning in UAE
Attended Core42 meeting in Abu Dhabi with Greg, Bjorn, Max, and Adam Jackson. Provided real-time technical assessment to the team via Slack, identifying product-customer fit across four CIQ product lines: RLC-H for defense customers (CVE remediation pain), Rocky as guest OS on Signature Cloud, Fuzzball as potential replacement for their unhappy AI cloud orchestration partner, and Ascender Pro for their heavy Ansible usage. Followed up personally with Raghu (EVP Engineering, Core42 US) offering in-person meetings.
Established 'what vs how' framework for Product-Engineering communication
In weekly sync with Brady and Brian, established a clear framework: Product defines the 'what' (exit criteria with specific, aggressive targets like 'ISO builds <4 hours'), Engineering defines the 'how'. Vague communication to leadership creates unnecessary reactive work and must stop. Escalation path defined: weekly syncs or on-demand to Peter for process breakdowns only.
Committed to creating CVE remediation value driver for GTM
Committed to creating a value driver for CVE remediation work after learning that remediation volume jumped from 1 to 86 per week. Timeline is ~2 months to develop the story after validating the new process is sustainable.
AMD RLC Plus Strategy: Speed-to-Market with Minimal Scope
Decided to prioritize speed-to-market for the RLC Plus AMD co-marketing launch. The initial build will use upstream AMD packages (pre-built ROCm), the kernel driver (not upstream DKMS), and enable EPEL. Deferring the more robust in-house rebuild until market traction is proven. Justin Haynes to draft proposal and decision matrix.
Time-based releases concept - trains leave on schedule
Consider moving to time-based releases where engineering ships whats ready on a fixed cadence (e.g., monthly or bi-weekly). Product must scope features to fit the timeline rather than engineering stretching to fit scope. The train leaves whether youre ready or not.
PRD first drafts are gravel - meant to be thrown away
Get product to understand that the first iteration of a PRD exists to be thrown away. Its gravel, not precious. Engineering questions should come fast and furious, and the document should go through massive churn. Pride of authorship must be eliminated.
Stop coaching product, move to SLAs
Stop trying to teach product managers (Brady, Brian, Dawson) how to do their jobs better. Instead, provide prescriptive SLAs - clear timelines and direct questions. If they dont like the dates, they can restructure their requirements. Leave it on the floor and walk away.
RESF crisis comms - single source of truth
In a RESF crisis, publish one official blog post as the central source of truth. All responses on social media (Hacker News, LWN, HPC forums) link back to that post. Do not engage in real-time debates. Own the traditional news cycle, not social media.
Project Shackleton - RESF contingency infrastructure
Build a parallel mirror of all RESF infrastructure in AWS (Git repos, Koji, vault/pub, Mattermost history) with goal of restoring Rocky Linux builds within two weeks if Lewis triggers his kill switch. Everything built with CDK and Ansible for repeatable deployment.
Servant leadership requires clear targets
Servant leadership and mentoring are fine, but must be paired with clear targets. Without clear targets, you cannot train the system. The AND between servant leadership and clear targets is mandatory - you cannot have one without the other.
Quality investment must serve velocity
Quality and automation investments are acceptable if the thesis is this will massively increase velocity in 3 months. Quality for its own sake is not the priority. Every quality investment should have a velocity payoff hypothesis attached.
Find the ceiling approach to velocity
Rather than incrementally improving 5% at a time safely, push until something breaks, then figure out if the breakage is fixable or a real ceiling. Air cover provided for aggressive experiments. Nobody gets fired for trying to go fast and breaking things.
Trustless processes over building trust
Focus on building contracts and processes that work without trust, not on building relationships. Good fences make good neighbors. Trust becomes a bonus, not a requirement. Contracts are what matter - relationships are nice to have.
PR standards in AI era - own the test suite, not the code
In an AI-enabled world, engineers should own the test suite and exit criteria, not necessarily every line of code. Quality comes from tests passing, not from reading every line. Engineer accountability shifts from I wrote this code to I own that this code passes these tests.
Team building over individual protection - Machiavellian approach
Stop protecting individuals at the expense of team success. The body being taken care of is the team, not individual engineers. Every day spent betting on someone who wont make it hurts the team. You already know who youre going to fire - just convince yourself youve done due diligence.
Trinity backfill for RLCAI/RLCH work, not maintenance
Use Trinitys backfill headcount to hire for RLCAI and RLCH work, not maintenance. Look for someone who can hold their own technically with Maple but will push AI adoption aggressively. Jamin identified as strong candidate - automation-first thinker, delivers and iterates, QA mindset.
Aggressive goal-setting philosophy - undercut estimates, force innovation
Set targets that seem impossible (e.g., 2 months instead of historical 6 months) and let the team figure out how. Success is not just hitting the target - its learning and attempting new approaches. The managers job used to be to pad estimates; now its to undercut them.
AI ownership standard - own everything you submit
Engineers must fully own everything in documents and code they submit, regardless of whether AI generated it. Using AI is expected and assumed. Submitting AI output you dont understand or endorse is not acceptable. Quality and accountability matter, not authorship.
Nathan job redefined - build a team, not deliver outputs
Nathan deliverable is a team that can adapt and learn, not technical outputs. Focus shifts from figure out how to deliver X with current team to build a team that can deliver what CIQ needs. This is fundamentally different from traditional engineering management.
Transfer Depot operations from Justin team to Steve team
Depot operations will transfer from Justin team to Steve team. Justin team will define the architecture for moving Depot to object storage, then hand off execution to Steve team for provisioning, infrastructure, and monitoring.
Address Damen visibility/recognition gap through demos
Damen will present at the Monday demo day (11 AM PST, on Bazel) to increase visibility. Peter will attend to show support. Peter committed to investigating and fixing the systemic visibility gap for all engineers, not just Damen specific case.
Pursue NVIDIA self-certification path for NVAIE integration
CIQ will pursue the self-certification path for NVIDIA NVAIE integration, targeting a GTC announcement. Will adopt NVIDIA preferred embedded license model, integrating NVAIE license cost (~$4,500/GPU) into CIQ product and providing L1/L2 support.
Related Patterns (8)
Executive Sponsorship for Strategic Partnerships
Strategic cross-company initiatives and major client partnerships require executive-level accountability to move at the right pace and ensure proper prioritization.
Small Circle for Sensitive Operations
When executing sensitive strategic operations, keep the circle of informed people as small as possible to prevent leaks that could accelerate hostile action or undermine the initiative.
Protect Engineering Capacity
When external demands threaten to overload engineering capacity, protect capacity by either requiring the demand to come with additional resources, or forcing hard prioritization choices upstream.
Proactive Talent Pipeline Investment
Invest in building leadership bench and talent relationships before there is an urgent need. Use proven relationships from past experience to create optionality.
Lead by Example with New Tools
When championing new tools or processes, personally use them and share results rather than just advocating. Learning by doing and demonstrating value through example is more effective than mandates.
Accountability Follow-Through
When you issue a warning or mandate with stated consequences, you follow through. Warnings are not threats - they are commitments. The credibility of future accountability depends on following through now.
Protect Engineering Focus Through Process
When faced with requests that would disrupt engineering focus (from sales, governance, product, or other stakeholders), establish processes that protect engineering ability to innovate while still satisfying legitimate concerns. Prefer systematic solutions over ad-hoc responses.
Three-Lever Talent Management
When pursuing a velocity or performance mandate, simultaneously operate on all three talent levers — upgrade (hire better), retain (protect key people), and exit (remove blockers) — rather than sequentially. This creates compounding momentum: exits free capacity for upgrades, retention preserves institutional knowledge during transitions, and upgrades raise the performance bar that justifies further exits.