What a standing bench actually is
Every federal buyer eventually asks the same question, and it is the right question: who is going to do this work? The common answer is a stack of resumes pulled from a recruiting database, attached to a proposal, belonging to people who have never heard of the bid. We run a different model. Precision Federal keeps a standing bench of engineers, licensed professional engineers, and domain specialists we have already scoped with, already agreed a rate basis with, and who have already told us in writing that they are willing to be named. The bench is organized by domain, because that is how the work arrives: defense, health, energy, transportation, and public-sector data.
A bench is not a payroll and it is not a recruiting pipeline. It sits between the two. Each person on it has a defined domain, a defined kind of seat they would fill, a rate we have agreed, and a written record of consent. Nobody sits idle waiting for us. They keep their own practice, their own faculty appointment, their own consulting book. What they have given us is a standing answer to one question: if a bid in your domain comes up, may we name you, and on what terms.
That single agreement, made months before any solicitation exists, is what lets a bid go out with a real human being in every seat instead of a job description. Our own rule is absolute and it costs us pages sometimes: no unnamed personnel. If we cannot put a consenting human name against a role, the role comes out of the proposal. A seat labeled "Senior Clinical Informaticist (TBH)" is a promise to hire, and evaluators read it exactly that way.
Bench Depth by Domain — Where We Can Name Someone Fast
Internal editorial weighting of how quickly we can name a consenting specialist in each domain. Illustrative, not a measured statistic.
Why the resume farm fails everyone in the room
The resume-farm model is simple. A recruiter builds a database. A capture team pulls three CVs that match the labor category. The names go in the proposal. If the bid wins, somebody starts calling. This is legal, it is common, and it fails three different people at once.
It fails the government, because the evaluated proposal and the delivered team are different teams. The Government Accountability Office has a whole line of protest decisions on this, usually filed under "bait and switch." The standard is narrow and hard to meet: a protester generally has to show the awardee represented personnel it did not reasonably expect to be available, that it knew or should have known that, and that the agency relied on the representation. Narrow does not mean harmless. Contracting officers have seen the pattern often enough that a proposal full of unfamiliar names now invites the question directly, and many solicitations answer it in advance with a key-personnel clause requiring written approval before any substitution.
It fails the specialist, who finds out that their CV was in a losing proposal for a program they would have loved, or in a winning one for work they cannot take. Nobody asked. Nobody paid. The specialist's name carried risk in a competition they were not party to, which is a real thing to do to a licensed engineer or a practicing clinician whose professional reputation is the asset.
And it fails the buyer, who is left holding a staffing risk they did not price. A prime that adds a subcontractor for a critical technical area is going to be judged on that area. Under FAR 15.305(a)(2)(ii) a contracting officer may consider the past performance of subcontractors that will perform major or critical aspects of the requirement. That is a real advantage when the sub is real. It is a real exposure when the sub's team is a database query.
Engaged by domain, not by headcount
The unit of our bench is a domain, not a labor category. "Senior Data Scientist" tells a buyer nothing about whether the person has ever seen a hospital's HL7 feed, a utility's SCADA historian, or a state DOT's crash file. So the bench is indexed by the thing that actually transfers: what problem has this person lived inside, in what institution, under what regulator.
That indexing is why our proposals can put an unusual pairing on a single scope. Federal AI and ML work almost always has a second discipline hiding in it. A grid-analytics build needs somebody who has run a distribution feeder model. A clinical decision-support build needs somebody who has actually signed a note. A structural monitoring build needs a professional engineer licensed by a state board, because at some point a stamp is going to matter and no amount of model accuracy substitutes for it. We carry those people because the work carries them.
The engineering core is ours and stays ours. Our team builds the production AI, ML, data, and cloud systems, and our team is led by a former professor in technology who ranks in the Kaggle Top 200 of more than 200,000 competitors and holds seven cloud certifications, with twenty years of building production systems for federal agencies across five consulting firms, three of them federal. The bench specialists are not there to build the system. They are there so the system is built against how the work is really done.
Written consent before the name is printed
Nothing goes into a volume without the person's written agreement. Not a verbal yes, not an assumed yes from a conversation last quarter, not a LinkedIn profile. Written, dated, and specific to the kind of work.
Consent covers two things at once, and this is a distinction a lot of firms get wrong. Consent to be named is consent to the bio. If someone agrees to appear in our proposal, the bio we print says where they worked, what they built, and what they are licensed or certified to do, in plain terms an evaluator can check. Scrubbing an employer down to "a large defense integrator" or a school down to "a major research university" does not protect anyone. It just makes the bio unverifiable, and an unverifiable bio scores like an absent one.
Consent is also revocable, and the workflow assumes it. People take full-time roles. Faculty go on sabbatical. A clinician's institution changes its outside-activity policy. When that happens the name comes out and either a different named person goes in or the seat is deleted and the scope is rewritten. That is a real cost on a bid week and we pay it, because the alternative is printing a name that will not show up.
How a Seat Gets Filled — Scope to Kickoff
Paid from the contract when it funds
This is the part that makes the model honest. Bench specialists are paid out of the contract that names them, at the rate agreed before the bid, for the hours written into the proposal. Nobody is asked to work for free on speculation, and nobody is asked to accept a rate reduction after award because the budget got tight. The rate is set once, in writing, at the moment consent is given.
The federal cost rules make this the clean path anyway. Professional and consultant service costs are allowable under FAR 31.205-33 when there is a written agreement covering the nature and scope of the service and the rate basis, and when there is evidence the work was performed. Firms that skip the agreement and improvise the rate later discover this during an incurred-cost review, which is the worst possible time. Doing the paperwork before the bid is not extra work. It is the same work, done when it is cheap.
Structuring the bench as engaged specialists rather than leased staff also keeps our size status clean. Under 13 CFR 121.106, SBA counts as employees the individuals employed on a full-time, part-time, or other basis, including people obtained from a temporary employee agency, professional employer organization, or leasing concern. Independent domain consultants engaged under their own agreements are a different arrangement, and getting that distinction right up front avoids an ugly size-protest conversation on a bid we won.
The performance-of-work rules cut the same way. In SBIR, the small business must itself perform at least two-thirds of the research in Phase I and at least half in Phase II. In STTR, the small business performs at least 40 percent and the single research institution at least 30 percent. A bench of specialists sized to advise, validate, and review fits inside those floors without strain. A bench sized to do the building does not, and firms that staff that way find out at the negotiation, not at the bid.
| What the buyer is really asking | Resume farm | Standing bench |
|---|---|---|
| Who shows up at kickoff | Unknown until recruiting closes | The person printed in the volume |
| When the specialist was contacted | After award, sometimes after protest | Before the bid, in writing |
| What the bio can say | Generic, scrubbed, hard to verify | Named employers, real projects, real credentials |
| How the rate is set | Negotiated after award, under pressure | Agreed at consent, unchanged at award |
| If the person declines | Substitution request to the CO | Caught before submission; seat re-named or removed |
| Who carries pre-award risk | The buyer, unpriced | Us, deliberately |
What this gives a prime or an integrator
If you are a capture manager, the bench is a way to buy a technical area without buying a hiring problem. You get named people, with real bios, in time for your volume. You get a subcontractor whose past performance you may cite for the major or critical portions of the requirement under FAR 15.305(a)(2)(ii). And if your contract carries a subcontracting plan under FAR 52.219-9, which applies to contracts expected to exceed $750,000 with subcontracting possibilities, the dollars flow to a small business that is doing engineering rather than passing through hours.
The teaming mechanics are ordinary and we keep them ordinary. Non-disclosure first, then a teaming agreement that says who does what work, then a subcontract at award. Where consent to subcontract applies under FAR 52.244-2, we give you the documentation package in the form the contracting officer wants it, not a week after they ask. On SBIR and STTR work, data rights get settled in writing before anything is built, with SBIR-developed items marked under DFARS 252.227-7018 and the protection period running twenty years from award under the SBIR Policy Directive. Nobody should be discovering a rights question during delivery.
What this gives a university PI
STTR requires a small business partner and a research institution, and the split is set by rule: at least 40 percent of the work to the small business, at least 30 percent to the institution. What is not set by rule is whether the small business shows up with a real team. Faculty who have been through a bad STTR know the failure mode, which is a partner that treats the university as a compliance checkbox and then disappears into its own scope.
The bench changes that conversation because the names are available before the proposal deadline, not after. A PI can see who would be building the software, who would be handling the cloud and security posture, and which domain specialists would sit between the research and the fielded system. Our team handles the paperwork side that eats faculty time, including the allocation-of-rights agreement, the budget build, and the registrations, and the institution's technology transfer office gets a counterparty that already knows what it will ask for.
What this gives a specialist
If you are the person on the other side of this, the offer is plain. You are named with your permission and never without it. Your bio says what you actually did, at the employers you actually worked for, because that is what makes it worth printing. The rate is agreed before your name goes anywhere. You are paid from the contract when it funds, for the hours in the proposal, and you are never asked to write unpaid proposal content in exchange for a maybe.
What we ask in return is small. Tell us the domains you want to be considered in. Tell us the kinds of work you would say yes to and the kinds you would not. Answer within a couple of days when a bid comes up, because bid weeks are short. And tell us when your availability changes so we can take your name out before it costs anyone anything.
Domains where we are actively adding named specialists
Clinical informatics and health data (HL7, FHIR, claims). Grid and utility operations data. Rail, transit, and state DOT data systems. Cybersecurity and RMF assessment, including cleared practitioners. Licensed professional engineers in civil, structural, and systems disciplines. Test and evaluation for autonomous and safety-critical systems.
The rules we hold ourselves to
- No unnamed personnel in any volume. A consenting human name, or the seat is deleted and the scope rewritten.
- Written consent before the name is typed, covering both the naming and the exact bio text.
- Employers and credentials named, never scrubbed. An unverifiable bio is worth less than no bio.
- Rate and hours agreed before the bid, and unchanged after award.
- Paid from the contract that names them, on the hours in the proposal.
- No exclusivity demanded for a bid that has not been won. Specialists keep their own practice.
- Availability re-confirmed before every submission, not assumed from the last one.
What to send us
Two different readers get here, so two different asks, both small.
If you are a buyer and you want to know who would actually be on your project, send the scope in a paragraph, the domain, and the date you need an answer by. You will get back the real names we would put against it, the real bios, and the hours we would commit, within one business day.
If you are a specialist, send a one-page CV and the two or three domains you want to be considered in. You will get a real answer within two business days, and if there is a fit we will tell you what kind of work we would name you for and on what rate basis before you commit to anything at all.
Frequently asked questions
No. A staffing arrangement sells hours against a labor category and the buyer directs the work. A bench is engaged by domain to advise, validate, and review a scope we are responsible for delivering. The distinction matters for cost treatment under FAR 31.205-33 and for size status under 13 CFR 121.106, and it matters more for whether the work is any good.
Most federal contracts carry a key-personnel clause requiring the contracting officer's written approval before substitution, and we follow it. Practically, we re-confirm availability before every submission so the problem is caught pre-award, and we propose a replacement of equal or better qualification with the bio and rate documented the same way the original was.
Yes. Contracts expected to exceed $750,000 with subcontracting possibilities generally require a subcontracting plan under FAR 52.219-9, and dollars subcontracted to a small business concern count toward the plan's goals. We are a small business under NAICS 541715 and 541512, SAM.gov active, CAGE 1AYQ0.
The small business must perform at least two-thirds of the Phase I research and at least half of Phase II, so outside specialists fit within the remaining share. In STTR the floors are 40 percent to the small business and 30 percent to the research institution. We size the bench to the advisory, validation, and review work that fits inside those limits.
Not for a bid that has not been won. Exclusivity before award asks someone to give up other opportunities in exchange for a probability, which is a bad trade for them and produces a resentful teammate for us. Post-award scope exclusivity on a specific program is a normal subcontract term and is negotiated then.