RecMaf is an AI-native recruiting platform — AI agents working alongside 340+ specialist recruiters in industrial manufacturing, construction and engineering. You post a role. We source, screen, score and verify. You get a ranked shortlist and pick who to interview. We close the rest.
This is ground we already work — our recruiters place roles into PEB, building materials and industrial manufacturing today, alongside some of the largest companies in India.
Magicrete — India's largest AAC blocks and building-materials manufacturer — hires through RecMaf today. That's the same construction-materials, plant-and-site hiring motion EPACK Prefab's PEB business runs on.
We've scoped this proposal to the two areas at epack.in that matter most right now: the core PEB business — design, manufacture and erection of pre-engineered steel buildings and sandwich panels, ~90.6% of revenue — and the data centre vertical growing fast inside it. Each is, underneath, a stream of engineering, plant and project roles that have to be found and placed fast enough to keep pace with commissioning dates and order deadlines. Here's how we read the direction, and the kind of individuals each front needs. You tell us the exact openings; we close them.
Turnkey design, manufacture and erection of pre-engineered steel buildings and insulated sandwich panels, serving 1,200+ customers across industrial, commercial and infrastructure segments. Mambattu Phase II went commercial on 29 April 2026, taking installed capacity to 147,122 MTPA across three plants — Greater Noida, Ghiloth and Mambattu. A fourth, a 50,000-tonne Gujarat greenfield plant, is under civil construction now, targeting commercial production by Q4 FY27. Every one of these sites needs its plant and project team in place ahead of the first shift.
Renewable, EV, semiconductor and data centre projects — the "new-age" sectors — now make up 35–38% of the order book as of May 2026, up from 25–28% in January. Management calls out strong, specific demand in data centres and battery storage within a ₹1,350 Cr order-visibility pipeline. These builds lean hard on EPACK's sandwich-panel and clean-room-grade enclosure work, and tend to bring larger, more complex, repeat-order relationships as hyperscale and colocation customers expand.
Each front is a distinct hiring stream with its own profile and its own geography. A generalist consultant treats them as one list. We treat them as two — with specialist recruiters who already live in each. We don't ask you to confirm how many roles you have; you bring the openings, and the platform aims the right specialists at each one.
A representative EPACK Prefab requirement — Design Engineer (Structural), Data Centre & Clean-room Enclosure vertical, Greater Noida — moving through the platform end to end. Every stage below runs today.
The marketplace, the AI models and the verification all run underneath. From your side, the process never changes.
Title, plant or site location, CTC, what good looks like. Five minutes — no briefing calls, no back-and-forth.
AI agents activate specialist recruiters across geographies. Every profile sourced, screened, scored and verified before it reaches you.
Review a ranked shortlist, choose who to interview. We handle scheduling, follow-ups, offer alignment and the admin.
Six things come together in one service. Each one replaces something your TA team currently stitches together by hand — across vendors, spreadsheets and follow-up calls.
Every opening is worked by multiple industrial & engineering recruiters at once — not one consultant who quietly parks the hard roles. They compete; the one who brings the candidate who actually joins is the one who earns it.
Every profile is scored against your JD on fifteen parameters and verified by an AI call before it reaches you. Your team reviews a ranked shortlist with reasons attached — never a pile of unscreened CVs.
Scheduling, reminders, candidate prep, feedback loops and offer alignment — the chasing that eats a large share of a TA team's week, across every plant and site. We run it on the platform.
Every JD, role nuance and hiring-manager preference — from a Mambattu shift-incharge brief to a data-centre-vertical design role — is captured once and reused on every future opening.
Instead of onboarding a separate consultant for Greater Noida, Mambattu, Ghiloth and the new Gujarat site, a single platform reaches every cluster PAN India — where the specialist recruiters already live and work.
No platform fee and no retainer. You pay a flat 5% of the candidate's CTC, only when an opening is actually filled — and a 3-month free replacement is built into every placement.
Everything you watched in the close runs on one engine, in production today. It sits between the recruiter and your HR team — reading every profile against your role, ranking it on fifteen parameters, calling the candidate to verify the claims, and keeping every status in one place. Here's each part, and what it actually does.
The moment a recruiter submits a candidate, the model reads the CV against your exact JD — not keywords, but role fit, how deep the plant or design experience runs, seniority and domain match. Weak matches are filtered at source, so they never reach your team's inbox.
Fit alone isn't enough to make an offer. The gate runs fifteen checks across three things a CV won't tell you — whether the person is genuinely available, whether the claims are real, and whether the profile holds together — and collapses them into one ranked score.
A score can be gamed; a conversation is harder to fake. After submission, the agent calls the candidate and checks the things that actually decide a hire — true notice period, real plant or design-tool experience, whether data centre / clean-room exposure is genuine — then flags any gap before it reaches your HR team.
The engine has a conversational layer on both sides. Recruiters get an instant read on a candidate before they submit. Your HR team asks the pipeline questions in plain language on WhatsApp — no dashboard, no login — and gets the answer from the same live data the platform runs on.
When an opening goes live, the engine decides which specialist recruiters to put on it — by domain, by track record on similar roles, and by geography — and brings more online automatically if the role stays open.
Everything lands in one place — a Kanban board with every candidate's stage, score, recruiter note and live status across every plant and site. Accountability runs past the offer letter: a portion of each placement is held for 90 days and released only once the hire has stuck.
| Factor | Traditional consultant | RecMaf |
|---|---|---|
| Industry depth | Generalist across sectors | Specialist bench in engineering & industrial manufacturing |
| Recruiters per role | 1–2 assigned | Multiple specialists, in parallel |
| Hard-to-fill roles | Quietly deprioritised | Incentive escalates within your range — more attention, not less |
| First profiles | 5–10 days | Days, not a week or two — on most plant & site roles |
| Screening | Manual, inconsistent | AI-scored & verified against the JD |
| Fee | 8.33%–16.67% of CTC | Flat 5% of CTC — below the market floor |
| Commercial model | Retainers, advances, complex terms | Success-fee only — invoiced on joining, payable in 15–20 days |
| Visibility | Black box until CVs appear | Every profile, score and stage live |
| Post-joining | Weak 30–60 day guarantees | 3-month free replacement, no charge |
There's no platform fee, no retainer and no subscription. Our fee is a flat 5% of the hired candidate's annual CTC — roughly a third of what a traditional consultant charges — and it's success-based: we invoice only when the candidate joins, with payment due within 15–20 days. Every placement carries a 3-month free replacement. RecMaf handles everything underneath — including paying out the recruiter network — so you deal with one partner and one number. While we're still building the product, all we ask in return is your team's insight on what else we can take off their plate.
One number for every role: 5% of the hired candidate's annual CTC. No per-role negotiation, no retainer — and well below the 8.33%–16.67% a traditional consultant charges.
Before a single role goes live, we sit with the people who are actually hiring — plant heads, project leads, HR — learn your openings end to end, and load everything into the platform, so your team never has to coordinate it again.
We meet the team that's hiring — TA and the plant / project heads — and understand how you work and what good looks like.
Your hiring team is set up on RecMaf. One place for every role, every profile and every status, across every site.
Together we map every opening — how many, which plant or project, the hiring manager for each, and how we'll close them.
The Gujarat plant build-out and the data centre vertical's growing order pipeline get first attention — the roles with real dates and real revenue behind them.
Every JD, every role nuance, every hiring-manager preference and site detail — captured once, in one place. From then on, recruiters source against it and the AI screens against it. You stop re-explaining roles and coordinating between consultants. That time comes back to your team.
Thirty minutes with your hiring team to understand your open roles across Mambattu, Ghiloth, Gujarat and Greater Noida — and to show how the platform takes the coordination off your desk. Share this with whoever should be in the room.
Scoped to the PEB and data centre businesses at epack.in. Timelines, incentive escalation and escrow splits shown are illustrative. Role counts, capacities and figures are indicative — compiled from EPACK Prefab's FY26 results, stock-exchange disclosures and public reporting — and intended for discussion.