A small partnership doing the unglamorous work of applied AI.
Fulcrum AI exists because most operators who could benefit from AI right now can't justify the cost or risk of a full internal team — and most vendors aren't going to meet them where they actually operate. We sit in that gap. We build the system, we own the outcome, and we leave behind something the customer can run.
Why Fulcrum, why now
The model layer has commoditized faster than anyone predicted. What hasn't commoditized: the engineering discipline to turn a model into a system that runs in production against real operating data, integrates with real workflows, and survives contact with the people who have to use it.
That gap — between "the model can do this" and "we ship this" — is where Fulcrum operates. We don't compete with foundation labs. We don't compete with hyperscaler consulting arms. We compete with the status quo of "we'll get to AI next year."
The practical difference shows up in what we refuse to skip. Most AI vendors treat governance and cost control as things you figure out after the pilot works — a compliance doc and a usage dashboard bolted on once a committee or a finance lead asks. We build both into the agent layer from the first week, alongside an assessment of the data underneath, because a system that can't be audited, can't be stopped, and can't be costed isn't production-ready no matter how good the demo looked.
How Fulcrum and DCS fit together
The build side. Engineering-led applied-AI partnerships: agent + MCP architecture, real-system integration, eval and observability, end-to-end delivery. Fulcrum is the prime entity on the contract.
The channel and advisory wing of Fulcrum. DCS reaches operators who would never source an applied-AI partner through the normal vendor process — and surfaces the partnerships where the build side can actually help.
DCS finds the engagement and frames the operating problem. Fulcrum scopes, builds, and ships. Together they cover the full surface area of how an operating company actually adopts AI — from "should we" to "this is running."
Who's behind it
Two co-founders. Narrow on purpose: every engagement gets a named technical owner, not a rotation through a bench.
Edward Harof
Co-Founder and CRO
Two-plus decades at the intersection of cloud platforms, channel partnerships, and AI go-to-market. Has led sales, marketing, and international partner alliances for cloud and digitalization software, and is active in major partner ecosystems — Microsoft Dynamics 365, Azure, iPaaS — that Fulcrum builds against today. At Fulcrum and DCS, sets the operating bar, owns the channel side, and carries the partner-side commitments behind every engagement.
Lok Acharya, PhD, PE, PMP
Co-Founder and CEO
PhD in Systems Engineering (Florida Institute of Technology), MS Electrical Engineering (George Washington University), licensed Professional Engineer and PMP. Leads Data Engineering and Solutions Innovation at an AI/BI managed-services firm with offices in Atlanta and Dallas — running data-engineering work, AI-readiness assessments, and the production execution of applied-AI initiatives across the BI platform. At Fulcrum, builds the agent, MCP, and integration layer behind every engagement. Python by default; Anthropic SDK and MCP servers for everything that has to run in production.
Operating norms partners can plan around.
- One technical owner per engagement. Partners get a named engineer accountable for what ships, not a tier of handoffs.
- Weekly written check-in. What shipped, what's stuck, what the eval shows. Same format every time so partner teams can plan around it.
- Narrow concurrent book. We work with a small number of partners at a time. Capacity is something we underwrite, not promise.
- Defaults to documentation. Every system we build ships with runbooks, eval suites, and architecture notes a successor team could pick up.
Currently working with…
A small set of build partnerships, an operator advisory engagement in a regulated vertical, and a channel relationship through DCS. Specifics are anonymized while engagements are live. See the work page for the shape of each.
If you're an operator, advisor, or platform team trying to figure out what to build, what to buy, or what to partner on — we'd like to talk.