Sciens Academy

The forward-deployed
AI engineer, trained.

A structured fellowship: two weeks of guided pre-work, one week of on-site residency, and six weeks of supervised practicum. Graduates earn the CSAI-FDE Associate credential, qualified to join a deployment pod or client platform team as a junior forward-deployment engineer under senior supervision.

9 weeks
320 contact hours
70B+
Open-weight models
1:8
Instructor to fellow cap
By invitation
Admissions

The Academy Charter

  1. We train engineers who can implement, deploy, and operate public, hybrid, and private AI where client data already lives.
  2. Taught on real hardware, under real constraints, inside real regulated environments.
  3. Certificates are earned through demonstrated deployment, never through attendance.
  4. The runbook belongs to the client team. Capability transfer is the objective.
Where the work happens

On your hardware.
Inside your controls.

Who we train

Built for the institutions that cannot afford to get AI wrong.

FINANCIAL SERVICES

Banks, asset managers, insurers

SR 11-7 · DORA · MAS · FCA. Deployed under model risk and third-party oversight.

HEALTHCARE & LIFE SCIENCES

Providers, payers, research

HIPAA · HITRUST · GxP. Deployed against PHI residency and audit obligations.

PUBLIC SECTOR & DEFENSE

Sovereign and mission systems

FedRAMP-aligned · Air-gapped · CJIS-class. Deployed without egress.

Programs

Three tracks. One outcome: production.

02 · OPERATE

Practitioner Track

For analysts, clinicians, and operations leaders using Sanctum daily.

  • 3 days · 6 sessions · applied labs
  • Analysts · clinicians · knowledge workers
  • Three production workflows, signed off
03 · GOVERN

Executive Briefing

For boards, executive committees, general counsel, and chief risk officers.

  • 1 day · in-person or private virtual
  • Board · C-suite · GC · CRO
  • Documented AI governance stance
Curriculum

Five modules. Taught on live systems, not slides.

Module 01

Platform deployment

Sizing, topology, install, TLS, and identity wiring.

Module 02

Model lifecycle

Selection, evaluation, safe-swap, and signed provenance.

Module 03

Retrieval & connectors

Ingestion, embeddings, hybrid search, ACL propagation.

Module 04

Security & SRE

Policy engine, observability, patching, DR drills.

Module 05

Workflow & handoff

Production capstone deployed with faculty sign-off.

Download the curriculum (PDF) →Full syllabus available under NDA.
Learning outcomes

What a graduate can do, unsupervised, on day one.

CSAI-FDE · DEPLOY

Forward Deployment Engineer

  • Deploy a 70B-class model on air-gapped hardware and pass a Level-2 security review without escalation.
  • Instrument model, retrieval, and identity paths with production observability and DR runbooks.
  • Execute a signed model safe-swap with documented provenance and rollback.
  • Own an incident from detection to written post-mortem inside the client's change process.
CSAI-P · OPERATE

Practitioner

  • Design and ship three production workflows against real business SLAs.
  • Evaluate model output against a written acceptance rubric and reject unsafe responses.
  • Escalate correctly under the client's model-risk and privacy policies.
CSAI-E · GOVERN

Executive

  • Chair an AI governance committee against NIST AI RMF and ISO 42001.
  • Approve or reject a production AI change with documented criteria.
  • Defend the deployment to regulators, auditors, and the board.
Fellowship structure

Nine weeks. Three phases. One production deployment.

PHASE 01 · WEEKS 1–2

Guided pre-work

Environment provisioning, model theory, security baselines, and a graded readiness assessment before residency.

  • ~40 hours · self-paced with weekly office hours
  • Readiness exam · pass mark required to attend
PHASE 03 · WEEKS 4–9

Supervised practicum

Fellows deploy Sanctum inside their own environment with weekly faculty reviews and a signed production capstone.

  • 6 weeks · weekly 1-hour faculty review
  • Capstone signed off by faculty and client sponsor
Prerequisites & assessment

Admissions are engineered, not marketed.

Prerequisite 01

Engineering baseline

Three or more years in platform, data, MLOps, or security engineering. Comfortable on Linux, containers, and Python.

Prerequisite 02

Executive sponsorship

Named sponsor at the VP+ level with a scoped production workload the fellow will deploy against.

Prerequisite 03

Readiness assessment

Graded pre-work exam on model risk, retrieval, and identity. Pass mark 80%. One retake permitted.

Assessment

Rubric, not a certificate mill

Daily lab sign-off, end-of-residency deployment exercise, and a signed production capstone. Failure to complete the capstone means no certification.

Cohort

Small, deliberate, English-language

Eight fellows per cohort. Delivered in English. On-site at a Sciens lab or, for enterprise cohorts, at the client site.

Assessment & certification

Transparent rubric. Written pass criteria. Named appeals.

Rubric

Five graded competencies

Architecture & deployment (25%), model lifecycle discipline (20%), retrieval & data governance (20%), security & operations (20%), capstone delivery (15%).

Pass mark

80% weighted, no dimension below 70%

A fellow cannot pass overall while failing any competency. One retake permitted within 60 days.

Examiners

Two independent graders

One lead instructor and one non-teaching examiner. Disagreements resolved by the Academy Dean.

Retake

One capstone retake within 90 days

Structured remediation plan, faculty office hours, and a second graded capstone attempt.

Appeals

Written, 14-day window

Appeals reviewed by the Dean and one external faculty member. Decisions documented and returned in writing.

Contact hours

320 hours · ~32 CEUs

40 instructor-led residency hours, 40 guided pre-work hours, and 240 supervised practicum hours. Eligible for approximately 60 CPE credits (ISACA / (ISC)² self-report). Formal accreditation pursuit under way for 2026.

Faculty

Taught by the engineers who ship it.

Delivery bench

Forward-deployed engineers.
What they teach is what they operate.

Platform architects & governance

Designers of the stack. Former counsel and risk officers who know the failure modes.

CREDENTIAL 01

Lead instructors

Minimum 8 years forward-deployment experience at tier-one financial, healthcare, or defense integrators. Active production engagements every quarter.

CREDENTIAL 02

Governance faculty

Former general counsel, chief risk officers, and model-risk leads from regulated institutions. Members of NIST AI RMF and ISO 42001 working groups.

CREDENTIAL 03

Clearances & ratios

Clearance-eligible faculty available for public-sector cohorts. Held to a 1:8 instructor-to-fellow cap with a named teaching assistant per cohort.

Practicum governance

Six weeks inside your environment, on written terms.

GOVERNANCE 01

Intellectual property

All artefacts, runbooks, and configurations produced during practicum are owned by the client. Sciens retains no rights to client data or deployment code.

GOVERNANCE 02

Escalation & SLA

Faculty on-call during practicum with a 4-business-hour response SLA. Any production-impacting incident routes to the client's change-management process, not the fellow.

GOVERNANCE 03

Access & controls

Fellows operate under the client's identity, least-privilege, and audit controls. No Sciens standing access to client systems. Background checks and NDAs mandatory.

Certification

A credential earned in production, not in a classroom.

Sciens issues a tiered credential. The Fellowship confers the Associate tier, qualifying a graduate to operate under senior supervision. The Professional tier is awarded only after two live, signed-off client deployments in the field.

CSAI-FDE·A

Associate · Forward Deployment Engineer

Awarded on Fellowship completion. Qualified to operate under senior FDE supervision on a Sciens pod or client platform team.

CSAI-FDE·P

Professional · Forward Deployment Engineer

Awarded after two live, signed-off client deployments. Qualified to lead a workstream inside a regulated production environment.

CSAI-FDE·S

Senior · Forward Deployment Engineer

Conferred by the Academy Dean on portfolio review. Qualified to own end-to-end deployment across multiple regulated clients.

Calendar & investment

Four cohorts a year. Priced against outcomes.

Q4 2026

November cohort

Residency: Nov 9–13. Practicum through Dec 24. Applications close Oct 2.

Q1 2027

February cohort

Residency: Feb 15–19. Practicum through Apr 2. Applications close Jan 8.

Q2 2027

May cohort

Residency: May 10–14. Practicum through Jun 25. Applications close Apr 2.

Q3 2027

September cohort

Residency: Sep 13–17. Practicum through Oct 29. Applications close Aug 6.

Per seat

Fellowship investment

Per-fellow tuition disclosed on application. Includes pre-work, residency, practicum supervision, capstone review, and one retake.

Enterprise cohort

Private cohort of 8

Delivered on client site or in a Sciens lab. Pricing disclosed on scoping call. Includes governance briefing for the sponsoring executive team.

Reserve seats or scope a private cohort →Cancellation and transfer terms provided at enrolment.
Alumni outcomes

Graduates operating sovereign AI in production.

DEPLOYED

100% within 30 days

Every certified fellow to date has placed a production deployment into their sponsoring institution within 30 days of certification.

COHORTS

27 fellows · 6 institutions

Across banks, an asset manager, a national health system, and two public-sector agencies. Names withheld under client NDA.

RECERTIFICATION

94% annual recertification

Fellows remaining as named operators one year on. The residual 6% moved off the deployment for role changes, not for performance.

Post-academy support

Certification is the start of the relationship, not the end.

SUPPORT 01

Faculty office hours

Weekly for the first quarter after certification. Monthly thereafter, for the life of the deployment.

SUPPORT 02

Retake & remediation

One capstone retake included. Structured remediation plan for fellows who do not pass on first attempt.

SUPPORT 03

Recertification

Annual recertification against the current standards baseline. Required to remain a named operator on the deployment.

Standards we train against

Every module is mapped to the frameworks our clients are audited against.

NISTAI RMF 1.0ISO/IEC42001SR11-7EUAI ActSOC 2Type IIISO27001HITRUSTCSF
Build your FDE bench

Bring the Academy inside your walls.