Mammoth Growth send-off testimonials - 2026-05-05
Captured from a Mammoth Growth team channel screenshot the founder sent at 13:26 ET, his last day formally winding down the MG consulting engagement. Three colleagues spoke up; quotes are verbatim, attribution preserved. Original screenshot at
assets/2026-05-05-sendoff-screenshot.png.
Why this is in the vault
The MG engagement was the founder’s gating consulting income from 2022 through Q1 2026 (terminated cleanly per the 2026-04-23 settlement, governed by the ../../06-reference/internal-review-mg-harness-cc-wrapped-2026-04-13 harness review). Send-off testimonials are evidence of how senior data leaders inside MG actually felt about working with him - useful as ground truth for future RDCO positioning surfaces (about page, services pricing, MAC info-product credibility), and as a closing-chapter artifact worth preserving against the natural fade of personal memory.
Testimonials (verbatim, 13:08-13:09 ET)
Jess Morris (Mammoth Growth, 1:08 PM)
“Thank you for everything you’ve contributed to Mammoth, Ben! Wishing you nothing but the best in your exciting new opportunity.”
Sonya Kushkuley (Mammoth Growth, 1:09 PM)
“Thank you for everything Ben! It’s been great working with you all these years!”
Brady Smith (Mammoth Growth, 1:09 PM)
“Working with Ben has been an absolute pleasure. You will be missed!!!!! Best of luck in your next chapter”
Reading the signal
Three voluntary public-channel posts in a 1-minute window suggests genuine warmth, not performative team obligation - the cadence reads as colleagues actually noticing the moment and choosing to speak up rather than a coordinated farewell ritual. “All these years,” “absolute pleasure,” “you will be missed” track the multi-year-relationship shape, not the polite exit-interview shape.
Specific signals worth banking:
- Sonya: “all these years.” Multi-year relationship characterization. Sonya was on the engagement long enough to form a stable judgment.
- Brady: “absolute pleasure” + “you will be missed!!!!!” The five exclamation points are real warmth; this is not buttoned-up corporate-speak.
- Jess: “your exciting new opportunity.” Implies the founder communicated the phData transition in a positive frame; Jess responded supportively.
Meeting transcript (Let’s Mammoth, 2026-05-05, send-off portion 6:12-10:36)
Captured verbatim from the meeting recording transcript the founder shared at 13:44 ET. Dylan Cruise drove the deck. Tom Clinton (sounds senior-leader / partner) and Jeff Bartolini (same) added remarks. Founder replied at 9:04 and again at 10:28.
Dylan Cruise (6:12) - presenting the deck
Dylan is the load-bearing colleague to track here. He picked the framing, pulled the data, drove the conversation. His own words on each slide:
Critical caveat that reframes EVERY number on the deck:
“I do want to caveat all of these numbers are just from internal projects and repos that I personally have access to. So you should probably just times all these numbers by 10 for all the clients that I don’t have access to.”
This is load-bearing. The slide-1 BY THE NUMBERS figures (1.29M lines, 110 PRs, 6,866 hours) are MG-internal-repos only. Real total is roughly 12.9M lines, ~1,100 PRs, ~68,660 hours if Dylan’s 10x heuristic is right. The 6,866 hours of billable work IS likely accurate (Harvest data is centralized, not per-repo) but the code-volume numbers are conservative-by-10x.
On /check-board-equivalent volume:
“Ben has pushed almost more PRs in the last four months of this year than he did all of last year, which is an incredible amount last year. But it’s just cool to see that even down to his last minutes, Ben continues to build and push out code to improve our clients, to improve our internal repos. He just keeps on building.”
On the cc-wrapped platform (the strategic centerpiece):
“The most recent and impactful contribution I do want to highlight is that I think all of us on all of our projects right now are using the CC-wrapped repo and that multi-repo structure. So if you are familiar with this, this is where all of our cloud code skills and commands and workflows live inside this repo. And that was Ben’s brainchild. He built that from the ground up. He advocated for it internally. He got people trained up on it, on the projects that he was running, and it’s how we deliver our code every single day now. So it really did transform how we go to market and deliver for all of our clients.”
This is the strongest single positioning anchor in the entire send-off. MG’s lead architect on record stating that Ben architected, built, internally advocated for, trained the team on, and operationalized the agentic harness that “transformed how we go to market and deliver for all of our clients.” That sentence belongs in any future RDCO MAC / about-page testimonial - explicit permission asked of Dylan first.
On Nutrafol (4-year longest-client relationship):
“Just kind of breakdown of the different projects that he’s worked on, namely the Nutrafol one. It’s our longest running client. Ben’s longest running client. But he has truly transformed how that entire organization thinks and builds data. And I don’t mean that lightly. I mean, from the ground up, Ben’s initiatives there, the things that he drives, the workflows he drives has truly transformed Nutrafol as a business, as well as our business.”
The “I don’t mean that lightly” insertion is real. Dylan deliberately marked the claim as not-throwaway. “Transformed how an entire organization thinks and builds data” is a citable case-study anchor.
Closing framing:
“People love working with you. People love learning from you. So best of luck on your next adventure.”
Founder’s reply (9:04) - voice-archive worth banking
“Thank all so much. You know, the faces change, but even people have been here a while, like Jess and Sonya, you all are great. Or the newer faces coming on with these agentec processes, you know, Daniel, Brady, Unni, you all are all making an impact. It’s been great to get to know all of you all and work alongside you. Faces may change, but the work quality stays the same, and you all are going to do great.”
This is also a people-mapping disambiguator:
- MG veterans (multi-year): Jess Morris, Sonya Kushkuley
- Newer agentic-cohort (recent agentic-process work with the founder): Daniel Rodman, Brady Smith, Unni Nambudiripad
Tom Clinton (9:34) - senior leadership
“Ben, I don’t want to restate everything that Dylan stated, but your impact here is massive. And I just appreciate all your contributions and just the foundational work that you’ve put into all of the agentec processes that we’ve built, as well as your client leadership. You know, that shines through in every call that you lead and provide some technical detail on. So, you know, you’ll continue to grow and shine in your next role and wish you all the best as well. And the bend-shaped gap here is a pretty big one for us to fill. But, you know, the team here is obviously well up to speed and we’re working on it. So just appreciate all your contributions and best of luck for what’s next.”
Tom reads as senior-leader / partner. Two specific phrases:
- “foundational work… in all of the agentec processes that we’ve built” - confirms Dylan’s framing from a second senior voice.
- “the bend-shaped gap here is a pretty big one for us to fill” - the role-shaped-gap framing. Strong successional signal.
Jeff Bartolini (10:19) + Tom Clinton (10:25) - explicit door-open offer
Jeff: “Likewise. Thank you, man. And if the next thing’s not awesome, just email me.” Tom: “Doors open.” Jeff: “Doors open.” Founder: “I’ve got a history of boomeranging, so it’s always an option.”
This is a standing return offer from senior MG leadership. Worth banking explicitly: the founder has a no-friction return-path to MG if RDCO bets fail. That option is worth real career-risk insurance value. Two senior partners on the record extending it, founder verbally accepting the standing option.
Connection to RDCO surfaces
This content is load-bearing for future positioning in two specific ways:
-
MAC info-product / Sanity Check / RDCO about-page testimonials. When RDCO ships a paid offering, third-party social proof from data-savvy colleagues is materially more credible than self-claim. With these three colleagues’ permission, any of these quotes could appear on a “what people say about working with Ben” surface. Get permission first; don’t reproduce without explicit ask.
-
Network mapping for future RDCO bets. Jess, Sonya, and Brady are now part of the warm-network bench: data leaders who know the founder’s work directly. If RDCO ever needs ICP customer-discovery calls (e.g., for the bookstore-for-agents wedge that needs vertical-AI customer signal), these are first-call relationships, not cold outreach. Worth queueing them in the Notion Contact Candidates DB so they aren’t lost.
The send-off slide deck (Dylan-led, captured 13:34 ET)
Mammoth Growth assembled a slide deck for the send-off; Dylan led the conversation. Founder forwarded 5 slides via iMessage. Slides archived at assets/2026-05-05-sendoff-slide-N-*.png.
Slide 1 - CONTRIBUTION · THE CLIENT WORK
Numbers on the page:
- 11 clients. 24 projects. 6,866 hours. total span 2022 - 2026
- Featured pillar: Nutrafol. 4+ years, 6 SOWs, 3,378 hours. Mammoth’s longest single-client run.
- Hours by client (billable, Harvest + Productive 2022-2026):
| Client | Hours |
|---|---|
| Nutrafol | 3,378.4 |
| MSG Networks | 1,589.5 |
| Progress.com | 720.5 |
| ASAP | 335.5 |
| ecoATM | 280 |
| Vitamin Shoppe | 247 |
| Tubi | 155 |
- Recent era: AI Data SOWs at Progress.com, Vitamin Shoppe, ecoATM, Tubi.
This is concrete, durable career-archive data: the client list, the time-allocation distribution, the “Nutrafol was the biggest single relationship” fact, and the “AI Data SOWs are the recent-era shape” framing. The fact that Mammoth highlighted Nutrafol at 3,378 hours as the headline pillar, and called out four AI Data SOWs as the recent specialty, is itself a signal of how MG framed the founder’s contribution at exit.
Worth banking for:
- About-page / RDCO services positioning (concrete numbers > self-claim)
- MAC info-product credibility (real client roster across CPG, media, devops, e-commerce)
- Future ICP discovery (Nutrafol, MSG, Progress, ASAP, ecoATM, Vitamin Shoppe, Tubi are now first-call relationships across CPG, consumer healthcare, sports media, digital media, devops, fashion retail, and streaming)
Slides 2-5
5 slides total per the iMessage attachment count, but only slide 1 surfaced through the iMessage MCP attachment pipeline. TODO: founder to forward slides 2-5 individually so the channel tag can attach each. Once received, append rows to the table below.
| # | Topic | Captured |
|---|---|---|
| 1 | By the Numbers / “Ben’s run, in figures” (opening summary) | 2026-05-05 13:41 ET |
| 2 | Contribution / The Client Work | 2026-05-05 13:34 ET |
| 3 | Activity Over Time / “Going out hot” | 2026-05-05 13:37 ET |
| 4 | Contribution / The AI Stack | 2026-05-05 13:37 ET |
| 5 | ”Thank You, Ben” closing slide | 2026-05-05 13:38 ET |
(Note: founder forwarded slides out of order; one duplicate of slide 2 came through during the relay but the actual deck has 5 unique slides per the natural flow above.)
Slide 1 (deck-opener) - BY THE NUMBERS · “Ben’s run, in figures”
The bird’s-eye summary slide. Four big numbers in landscape layout:
- 4 years active - Jul 2022 - Apr 2026
- 110 pull requests - 87 merged, 79% merge rate
- 1.29M lines touched - across 1,388 files
- 6,866 billable hours - 11 clients, 24 SOWs
Source: GitHub REST/Search API + Harvest + Productive (Snowflake DBT_OPS_REPORTING_PROD).
This is the deck-opener: MG’s bird’s-eye framing of the founder’s full 4-year run before the deck zooms into client breakdown, time-acceleration, AI-stack architecture, and the closing thank-you.
The 79% merge rate is a notable signal: not every PR landed, but 4 in 5 did. Reads as “ships work that gets accepted,” not just “ships volume.”
The 1.29M lines / 1,388 files breakdown also reframes the AI-stack slide-3 number: ~78K lines of that total were the harness/plugin/prompt-library platform, meaning ~6% of the founder’s total touched-lines became durable internal-platform code that the rest of MG built on. The remaining 94% was client work. That ratio is a real engineering-leverage signal.
Worth banking for:
- Resume / about-page anchor. “4 years, 110 PRs at 79% merge rate, 1.29M lines, 6,866 billable hours, 11 clients, 24 SOWs.” Unimpeachable concrete numbers, MG-sourced.
- Engineering-leverage proof. 6% of touched lines became reused platform. That’s the kind of ratio MAC info-product can credibly claim teaches.
Slide 3 - CONTRIBUTION · THE AI STACK
Headline: “Architect of Mammoth’s cc-wrapped platform.”
Tagline: “Built it. Shipped it. Made it stick.”
Body: From scratch, Ben built the harness, plugin, and prompt library that the rest of us code with every day. Three repos, 78 PRs, 72 commits, ~78K lines of platform.
Bar chart - PRs and commits across MG-built AI repos:
| Repo | PRs | Commits |
|---|---|---|
| Repo 1 (largest) | 51 | 56 |
| Repo 2 | 19 | 17 |
| Repo 3 | 3 | 5 |
Source: PRs and commits across MG-built AI repos.
This is the most strategically important slide in the deck. It’s MG’s own framing of the founder’s load-bearing contribution: he built the agentic harness that the rest of the team uses. Three repos, ~78K lines of platform, and the team adopted it (the “made it stick” framing - not just shipped, but stuck). This is the directly-portable resume bullet for everything RDCO is building (Ray harness, MAC framework, agent-tooling skills, verify-action). The pattern shipping at MG is the same pattern shipping at RDCO.
Worth banking for:
- MAC info-product credibility. “I built the cc-wrapped platform that 11-person Mammoth Growth used every day, three repos, 78 PRs, 78K lines of platform code, and the team adopted it (made it stick).” Beats every other credentialing line.
- L5 north-star direct evidence. Building harness-grade agent tooling is the L5 thesis (../../06-reference/2026-05-04-karlmehta-llm-commoditization-intelligence-rails). Founder shipped it once at MG; doing it again at RDCO scale is the next step up.
- About-page concrete proof. Three repos, 78 PRs, ~78K lines, made it stick. Resume-ready.
Slide 5 - “Thank You, Ben” (closing slide)
Slide content (verbatim, from MG):
FROM THE WHOLE TEAM
Thank You, Ben.
You set the bar for what an Architect at Mammoth can be.
You’re an absolute beast.
Best wishes on what’s next.
- Mammoth Growth
Two specific phrases worth banking:
- “You set the bar for what an Architect at Mammoth can be.” Strong claim. Frames the founder as the reference standard for the role title.
- “You’re an absolute beast.” Italicized, set apart. Reads as deliberate emphasis, not throwaway.
Combined with the slide 3 framing (“Architect of Mammoth’s cc-wrapped platform”), this is MG canonizing the founder as the reference Architect. That’s a positioning anchor for any future RDCO surface that needs to claim “I architected agent tooling at production scale.”
Additional team testimonials (image 2, captured 13:38 ET)
5 more colleagues weighed in on the same channel a few minutes after the original 3-person batch. Image at assets/2026-05-05-sendoff-comments-2.png. Verbatim:
James Lee (1:09 PM)
“Ben, you’ll be missed!! Thanks for all your contributions and the learning opportunities. All the best for the future and hope to cross paths again!”
Founder reply (1:09 PM)
“Thank you all!”
Unni Nambudiripad (1:10 PM)
“+1 to all of the above - absolute pleasure working with you, I’ve learnt a lot from you, and I wish you all the very best!”
Nicole Griffin (1:10 PM)
“You’ll be missed! It’s been great working with you and having your guidance along the way.”
Daniel Rodman (1:11 PM)
“Wishing you well and the absolute best Ben! It was awesome to work with you!!”
Note: the “Well deserved! Congratulation Sawyer!” line in this image is the founder’s OWN reaction to a separate thread - MG promoted Sawyer to architect, and the founder posted to congratulate him. Founder confirmed 2026-05-05 13:45 ET: “It was not a move to backfill me. This was well deserved on his part. I’m really excited for Sawyer!” - so Sawyer’s promotion is concurrent with but not caused by the founder’s departure.
Reading the additional signal
- Total respondent count: 8 distinct colleagues (Jess Morris, Sonya Kushkuley, Brady Smith, James Lee, Unni Nambudiripad, Nicole Griffin, Daniel Rodman, plus the unidentified “Sawyer” comment if separate). Eight is a substantial team-channel response for an exit moment.
- Recurring themes: “absolute pleasure” (Brady, Unni), “learnt a lot from you” (Unni - mentor framing), “guidance along the way” (Nicole - mentor framing), “set the bar” (whole-team slide).
- The mentor-framing words from Unni and Nicole are the most strategically valuable for RDCO positioning - they explicitly describe a teaching relationship, which is exactly the MAC info-product pitch.
Slide 2 - ACTIVITY OVER TIME · “Going out hot”
Numbers on the page:
- 70 commits in the first four months of 2026. Headline number, big orange typography.
- “Already past his 2025 total - and he had eight more months to go.”
- Bar chart, commits per year (presumably 2022 = year 1 through 2026 = year 5):
| Year | Commits |
|---|---|
| 1 (2022) | 5 |
| 2 (2023) | 35 |
| 3 (2024) | 2 |
| 4 (2025) | 69 |
| 5 (2026, Jan-Apr 23) | 70 |
The shape: a near-zero year (2024 = 2 commits) followed by sustained acceleration into the AI-era work. 2026 already exceeds 2025 in a third of the calendar. MG’s framing reads as “he found his stride right as he was leaving.”
This connects to the recent-era AI Data SOWs framing from slide 1: Progress.com, Vitamin Shoppe, ecoATM, Tubi were all caused by the founder pivoting his contribution shape from analytics-engineering toward AI-data work. The acceleration is real and visible in commit signal, not just in revenue framing.
Worth banking for:
- MAC info-product positioning. “I went from 35 commits/year to a 70-commit-per-quarter cadence by adopting AI-tooling discipline” is exactly the credential MAC needs.
- L5 north-star validation. Per the L5 thesis (
~/rdco-vault/.../project_l5_north_star_strategic_direction.md), unhobbling the COO agent is the primary investment. Founder’s own commit acceleration is direct evidence that AI tooling scales individual output materially - the same dynamic RDCO claims. - About-page concrete number. “5x commit-cadence acceleration over 4 years driven by AI tooling adoption” beats “I’m bullish on AI” by every measure.
Action items
- (Founder, optional) Get permission from Jess / Sonya / Brady to use their quotes on any future RDCO marketing surface. One-line LinkedIn or Slack DM is plenty.
- Archive screenshot at
~/rdco-vault/01-projects/mammoth-growth/assets/2026-05-05-sendoff-screenshot.png(done) - File this testimonial note (done)
- (Ray) Queue Jess Morris, Sonya Kushkuley, Brady Smith to Notion Contact Candidates DB as warm-network contacts (done; 3 candidates queued).
- (Founder) Forward slides 2-5 individually so I can transcribe + archive them.
- Archive slide 1 at
assets/2026-05-05-sendoff-slide-1-client-work.png(done)
Related
- index - MG project root
- 2026-04-27-jeff-agentic-velocity-quantification-draft - other MG-era contribution
- ../../06-reference/internal-review-mg-harness-cc-wrapped-2026-04-13 - the agentic harness review built for MG
- ../phdata/career-transition - the phData transition this send-off references