Tech · Healthcare · USA · 8 min read

Healthcare SaaS saves ~$12,000 monthly

Managed remote team cuts monthly ops cost by $12k with full EOR, payroll, and compliance.

Monthly savings
$12,000
Annualised
$144k
Engineers managed
6
Compliance overhead
0
What they needed

The brief.

A US healthcare analytics SaaS, post-Series A, with a six-person US engineering team running hot and a runway that needed to stretch. The CEO wanted to scale to a 15-person team to ship three new product lines, but every additional US senior engineer was ~$240k loaded cost. A board observer suggested offshore; the CTO pushed back hard — past experiences had been senior-labeled juniors, time-zone chaos, and code that needed rewriting. They needed a model that delivered actual senior engineers with their existing leadership keeping technical direction, plus payroll, compliance, and hardware handled so the CTO didn't add HR to his workload. The shortlist would have to clear the existing US team's bar, judged the same way they'd judge a US hire.

Must-haves
  • Frontend: React.js + Next.js + GraphQL (production at scale)
  • Backend: Python + SQL + ETL (analytics pipelines)
  • HIPAA-aware development practices
  • Strong async written communication (CTO is async-first)
  • Daily standups in 9–11am PST overlap
  • References from prior US clients
Sourcing & screening

The funnel.

Sources: Internal database (40,761 candidates), LinkedIn Sales Navigator, Naukri, Indeed, AngelList Talent, niche dev communities (HN Who's Hiring, lobste.rs)

Applicants sourced
1,600
100.0%
AI-scored above 75/100
312
19.5%
Resume review by senior recruiter
88
5.5%
Curated profiles delivered
30
1.9%
Client shortlist
12
0.8%
Technical assessment
9
0.6%
Live coding (with US CTO)
7
0.4%
Cultural + reference
6
0.4%
Hired
6
0.4%
Challenge

The problem.

The CTO had a real and earned skepticism about offshore engineering teams. He'd inherited an offshore vendor relationship at his previous job that delivered code passing surface-level review but failing under load — N+1 queries everywhere, no test coverage, and an inability to debug production incidents because the engineers had been spun up on toy projects and never owned anything at scale. He'd vetoed the board's first offshore suggestion six months earlier. The CEO understood the cost math but didn't want to overrule the CTO's technical judgment, so the brief came to us with a stipulation: every candidate would face the same technical bar as a US hire, including live coding with the CTO. If we couldn't deliver senior engineers who held up under that pressure, the engagement would fail. There was a secondary challenge: the US team's culture was async-first, with heavy emphasis on written tech specs before any code got written. Offshore engineers from agency backgrounds typically come from synchronous, ticket-driven cultures. Cultural fit on this dimension was probably more important than raw stack expertise.

Solution

What we did.

We started with calibration, not sourcing. Two hours with the CTO and VP Engineering, reviewing past senior-engineer interview transcripts (with names redacted) to match our screen to their bar. We saw that they emphasized debugging stories and system-design judgment over algorithm puzzles, so our written assignment was a scoped real-codebase problem with intentional architectural traps. From there, Talent OS ran the funnel: 1,600 applicants pulled from the internal 40,761-candidate database plus four external sources, AI-scored against the calibrated rubric, narrowed to 88 by the engine, screened to 30 by our senior recruiter. Each of the 30 got the 4-hour async assignment. 12 cleared the bar and were presented to the CTO with full evaluation notes including a video of the candidate explaining their assignment solution. The CTO ran live coding sessions with 9 of the 12. Six engineers cleared his bar — all senior, all with 5+ years production experience, all with prior async-first culture exposure (verified through references). EOR contracts issued in five business days, Windows i7 laptops shipped to engineers' homes in India, Slack provisioned with role-based access, and we ran daily standups in the 9–11am PST overlap window. Talent OS productivity dashboards gave the CTO weekly velocity, code-review-turnaround, and incident-response visibility — surfaced any performance issues before they became friction in the team relationship.

Outcome

What changed.

Net monthly cost dropped $12,000 — about $144,000 annualized — across overhead, talent cost, and HR/management combined. All six engineers passed their 90-day probation. The CTO's initial skepticism converted within the first sprint when the senior-most placement (Bengaluru-based, 8 years experience) found and fixed a long-standing N+1 query that had been a recurring on-call page for the US team. Productivity rose: the team's sprint velocity went from a 36-story-point average to 58, and PR-cycle-time dropped 41% as the offshore engineers absorbed code review across the team. Compliance overhead from the client's perspective: zero. Hardware logistics: zero. Statutory filings: zero. Eighteen months on, four of the six engineers are still with the team. One was promoted to Tech Lead. The CTO recommended us to a peer at a competing healthtech, which became our second large US healthcare client. The CEO's runway extended by an effective five months versus the all-US scaling path she'd modeled at board.

Process

How we ran it.

01

Brief & calibration

Two-hour calibration call with the CTO + VP Engineering. Reviewed three past senior-engineer interview transcripts (anonymized) so our screen matched their bar. Locked must-haves vs nice-to-haves on a single page.

02

Sourcing & AI scoring

Talent OS pulled 1,600 applicants from the internal database and four external sources. AI scored on twelve dimensions (stack match, project complexity, employer signal, tenure). Top 88 reached human screen.

03

Curation + technical screen

30 candidates passed initial screen with our senior recruiter and got a written assignment (4-hour, async, scoped to the client's actual codebase patterns). 12 cleared the bar and reached the client.

04

Onboard + manage

EOR contracts in 5 business days. Windows i7 laptops shipped to all six engineers. Slack provisioned with role-based channels. Dedicated account manager runs daily standups, monthly perf reviews, and weekly client syncs.

Looking back

What made this work.

The whole engagement turned on calibration, not sourcing. Spending two hours with the client's CTO before any candidate was screened — and explicitly mapping our rubric to theirs — eliminated 80% of the typical 'this candidate is technically OK but not your culture' rejections that plague offshore engagements. Calibration is unglamorous and unbilled time, but it's the highest-leverage two hours we spend on any 6+ headcount engagement. The second lesson: dedicated account manager for managed-EOR engagements pays for itself in retention. Engineers who feel managed (weekly 1:1s, clear feedback loops, promotion paths) stay. Engineers who feel placed and forgotten leave within twelve months. We baked the AM into the per-seat fee so the client never had to choose between paying for talent and paying for talent retention.

Tech stack

What we built it with.

Frontend (existing client stack)
React.js + Next.js
Existing dashboards in Next.js 13 app router. Engineers needed prior production experience with the new patterns, not just CRA-era React.
GraphQL
Apollo Client on the frontend, Hasura on the backend. Engineers were tested on GraphQL schema design, not just consumption.
Backend (existing client stack)
Python
Service layer in FastAPI; data pipelines in pure Python. Engineers needed both — pipeline engineering is harder to hire for than API work.
SQL (Postgres)
Analytical queries against a 14TB Postgres instance. Tested on query optimization and partitioning strategy, not just CRUD.
ETL
dbt + Airflow + custom Python orchestrators. Calibration revealed the client cared most about dbt model design quality.
Talent OS (vetting platform)
AI scoring engine
Twelve-dimensional candidate scoring trained on past placement outcomes. Top 88 of 1,600 reached human review.
Written assignment harness
4-hour async assignment scoped to the client's actual codebase patterns. Auto-graded on structure; manually graded on judgment.
Reference framework
Structured reference calls with prior managers. Eight-question rubric on technical bar, communication style, and async-fit.
Ops (post-placement)
EOR contracts (India)
Full statutory employment via our India entity. PF, ESI, gratuity, TDS handled. Engineers got Indian employment terms with the client's intended cash comp.
Productivity insights
Talent OS dashboards on velocity, PR-cycle-time, and on-call response. Weekly digest to the CTO. No-surprise performance management.
Hardware logistics
Windows i7 laptops shipped to each engineer's home. Replacement SLA 7 days for hardware failures.
More cases

Related stories

Manufacturing · India
Manufacturing firm saves $2,600 on HRMS
46% off quote · 200 hours saved
Healthcare SaaS · USA
Non-tech founder launches Healthcare MVP
MVP to launch, zero code background