DISCOVER
I map the workflow end to end before I touch a tool. I sit in on calls, watch the actual work, and find the bottleneck most teams have stopped seeing. Most implementations fail here, not in the build.

I'm Sage Dutra. AI Solutions Architect.
I build systems that get adopted because they're designed around the people who use them.
I run discovery and a data-readiness check in week zero, then design for the workflow, not the demo.
I build for the people who have to use them, with hands-on training and support until it sticks. More than two hundred trained.
I attach one owned metric to one workflow before a line of code is written.
I scope agents conservatively and keep a human on the high-stakes call. I can also tell a real agent from a rebranded chatbot.
* Sourced from MIT Project NANDA, State of AI in Business 2025, McKinsey State of AI 2025, Menlo Ventures 2025, and Gartner 2025.





























































I've spent fifteen years building marketing, software, and operational systems for organizations that needed them to actually work.
The market stopped needing people who train models. It needs people who get bought tools adopted inside real work. I've been doing that for 15 years.
76% of enterprise AI use cases are now bought rather than built. Menlo Ventures, 2025.
The same method whether I am on staff full time or brought in for a project. Find the real bottleneck, design for the people who live with it, and stay until it sticks.
I map the workflow end to end before I touch a tool. I sit in on calls, watch the actual work, and find the bottleneck most teams have stopped seeing. Most implementations fail here, not in the build.
I translate what I found into a system on paper first: process maps, requirements, and the trade-offs stated plainly. The people who will use it review it before anything gets built.
I architect for adoption, not just function. Fault-tolerant, humans in the loop for high-stakes calls, and built to survive contact with real users instead of passing a demo.
I do not ship and disappear. Hands-on training, documentation, and post-launch support. More than two hundred people trained on systems I built, across thirty-plus organizations.
In a full-time seat this runs on a loop. On a shorter contract, the first two stages stand alone as a deliverable.
Each one proves something different. Reliability engineering, orchestration breadth, and multi-agent deliberation. The full deep-dive on each is one click away.
An autonomous, human-gated multi-agent pipeline that researches a local business, writes a custom site from scratch, runs outreach, and onboards the customer. Built solo, five iterations in five months.
Built for the failure mode that kills AI systems, the output that looks right and is wrong.
Production-grade reliability engineering, not a demo.
Production-grade, operated end to end on real data, pre-revenue by design.
An autonomous solution architect on LangGraph. Hand it a business process; it interviews you with an expert-persona panel, designs a corpus-grounded solution, then turns an adversarial panel loose until zero blockers remain, and emits a build-ready spec.
Orchestration breadth across two frameworks, plus adversarial-loop and cost-engineering judgment.
Local dev tool, verified on real runs, not a deployed product.
An autonomous strategy council of seven AI personas with structured voting, task extraction, and an Obsidian-backed institutional memory.
Designing multi-agent deliberation and durable institutional memory.
Built solo.
The track record the AI work is built on. Salesforce deployments, a regulated intake platform, launches, and a department built from scratch. Each one led by the hardest number it can defend.

Built the entire go-to-market engine for a pre-revenue solar-hardware startup: brand, website, paid acquisition, sales videos, and a custom Salesforce instance. The pipeline I architected grew from $0 to $29M in twelve months, and the CRM scaled from 3 to 6 users as the team grew.
“Sage came recommended to us and we hired him for an off-site project on a short timeline. He was exceptional in his creativity, professionalism, and planning. He worked around the clock and through multiple iterations to give us exactly what we needed for our customers and investors.”

Ran the full public launch of a Portland gig-economy startup: positioning, paid social, two-audience email automation, SMS, website, and a custom Salesforce instance. The email system beat cited industry benchmarks by 3 to 5 times, a 33.6% open and 12.8% click against 17.9% and 2.3% averages, and paid acquisition landed registrations at $4.11 each.
“Sage is a highly intelligent and enthusiastic member of my marketing team. Success often depends on the people you surround yourself with. Sage's presence on the team increases my confidence we will be successful.”

Pioneered and led a 15-person department from scratch. Built the recruiting CRM, the admissions-approval workflow, and a content-analytics loop, then personally ran recruiting for the first months to refine the system before hiring. Wrote the manuals and onboarding, trained the team to own their sub-systems, drove cost-per-lead down to roughly $5 to $7, and built the outbound system that generated more than 60% of all applications. In a single year, enrollment grew from 15 to 118 students, a 47% applicant-to-enrolled conversion.
“One of Sage’s strengths is his ability to be incredibly focused and goal oriented while being genuinely interested in his clients. He wants the best solution possible for you and your team.”

End-to-end Salesforce deployment for a global nonprofit. Workflow automation across multiple countries, followed by six months of weekly hands-on training so every team actually adopted it. One of nine from-scratch CRM builds in the portfolio.
“Sage expressed an enthusiastic and encouraging approach to assessing our needs, provided logical customization after listening intently to our wishes and followed it all up with well-designed practical training to help us get up and running.”

Nine Salesforce implementations, seven built from scratch and two overhauls, across nonprofits and startups. The largest single migration moved more than 20,000 records from Highrise into Salesforce Nonprofit over roughly two months, with me as project manager and architect. Each from-scratch build ran discovery to adoption in about three months on an enforced maintenance-and-training retainer.
“Adept at not only the technical aspects of the work, but clear in communicating with those who were new to the platform or level of technology.”
Designed and shipped a solo-built, HIPAA-compliant digital intake system for a newly regulated field within a week of the regulations publishing. Consolidated more than 12 required forms into a 3-step process, moved sensitive data off unsecured email, and cut intake turnaround from 7 to 14 days down to a single day. Deployed to roughly 15 practices.
64 public reviews across two Google Business Profiles, Upswing and ClearHaven, at a 4.9 average.
"Sage came recommended to us and we hired him for an off-site project on a short timeline. He was exceptional in his creativity, professionalism, and planning. He worked around the clock and through multiple iterations to give us exactly what we needed for our customers and investors."
Across a decade of work and two companies, 64 public reviews at a 4.9 average. Upswing → · ClearHaven →
Heads up: you’ll see “Ryan” on older reviews.
I legally changed my name in 2025.
"A creative thinker with the ability to execute."
Star average across 64 public reviews, a decade of work.
"The area's premier marketing strategist."
"He did his preparation excellently, followed through on what he promised."
"Incredibly knowledgeable and easy to work with."
"A creative thinker with the ability to execute."
Star average across 64 public reviews, a decade of work.
"The area's premier marketing strategist."
"He did his preparation excellently, followed through on what he promised."
"Incredibly knowledgeable and easy to work with."
Years building systems clients still use today.
"Brings fresh ideas and energy to the table."
"He wants the best solution possible for you and your team."
"Well-designed practical training to help us get up and running."
"Intensely creative, with a knack for important details."
Years building systems clients still use today.
"Brings fresh ideas and energy to the table."
"He wants the best solution possible for you and your team."
"Well-designed practical training to help us get up and running."
"Intensely creative, with a knack for important details."
What I've built with, deeply enough to architect, deploy, and train on.
The things people usually ask first.
Because the data favors it. MIT found expert-partnered builds reach production at 67% against 22% for internal-only. The bottleneck is rarely engineering talent. It is discovery, workflow redesign, evaluation, and adoption, which is the work I specialize in.
No. What I have is fifteen years of building systems for thirty-plus organizations, with a 4.9 rating across 64 public reviews over a decade. Companies that have dropped degree requirements (Google, Apple, IBM, most AI-native startups in 2026) recognize that applied work is the better signal. If a four-year degree is a hard requirement, I'm not the right candidate. If results are, I am.
Not as a W-2 employee. I've built and implemented systems for thirty-plus organizations as a founder and consultant, including Salesforce deployments, custom intake platforms, and FORGE, an autonomous AI pipeline that runs end to end. I bring the perspective of having been the customer who had to make the technology actually work. That perspective is hard to develop from inside the industry.
Yes, on the side. Upswing is a productized website business for local trades, built on FORGE, an autonomous pipeline. Launch is intentionally held while I build supporting tools and run my job search, so it stays low-touch and does not compete with a senior W-2 role. The work I'm seeking is broader, faster-moving, and operates at a scale I cannot reach alone.
Immediately. I'm based in Salem, Oregon, fully remote-ready, and not in a notice period.
I've designed and built three AI systems solo. FORGE is the centerpiece: an autonomous multi-agent pipeline using Claude Max (Sonnet and Opus) for creative work and Gemini Flash for analytics, orchestrated with Temporal workflows for durable execution. It researches local businesses, writes custom HTML websites from scratch, generates personalized outreach, and carries built-in payment and a retention loop. Postgres for state, Vercel for deployment, per-prospect workspaces for context. Built solo across five iterations, with deterministic post-processing and a silent-homogenization detector doing the reliability work.
Alongside FORGE I built Founders Circle (an autonomous strategy council of seven AI personas with voting, task extraction, and an Obsidian-backed institutional memory) and the Solution Factory (an autonomous solution architect on LangGraph that interviews you with an expert-persona panel, designs a corpus-grounded solution, then turns an adversarial panel loose on the design through a bounded revision loop until zero blockers remain, emitting a build-ready spec behind two human gates). The deeper breakdowns are linked from the "Other AI Systems" section above.
Fully remote is the priority. I have five children with a 50/50 custody arrangement, which makes relocation impractical. For the right team and a strong remote culture, occasional travel for offsites and key meetings is fine.
More questions? Email sage.dutra@gmail.com.