Senior ML Engineer

About Evercred

Evercred is transforming physician credentialing from a 90–120 day manual process into a ~14-30 day automated workflow using AI agents. We are building a production system that orchestrates multi-step verification across state medical boards, education institutions, and employers—with real-time monitoring, intelligent escalation, and compliant audit trails.

Stage: Pre-revenue, 7 signed LOIs (~$250K), targeting $1M ARR by Summer 2026

Stack: Next.js, TypeScript, PostgreSQL, Prisma, Anthropic Claude API, HashiCorp Vault, LangChain-style orchestration

​

The Role

This is a fractional CTO-level individual contributor role focused on architecture, orchestration, observability, and reliability of a multi-agent verification system.

We are working with an external agency on agent implementation and need a senior in-house engineer to:

• Establish production monitoring and observability
• Review architecture for security and scalability
• Optimize agent performance and cost
• Build orchestration, decision logic, and escalation systems
• Ensure readiness for scale (25K+ physician wallets, 150+ enterprise customers)

Why this is interesting

• First production AI agent system in healthcare credentialing
• High-impact problem: eliminating 90–120 day bottlenecks that cost hospitals millions
• True multi-agent orchestration (parallel workflows, failure modes, latency variance)
• Non-deterministic systems engineering on LLM infrastructure
• Regulated environment: HIPAA-adjacent, Joint Commission auditability, long-term retention

​

What You’ll Do

Phase 1: Monitoring & Architecture Review

(Weeks 1–6 ½-time | 1–3 weeks full-time)

Observability & reliability

• Implement monitoring for agent workflows (performance, cost, success/failure)
• Build dashboards, alerting, and escalation detection
• Create debugging tools for non-deterministic agent behavior

Security & compliance

• Review architecture for sensitive PII handling (SSN, credentials, portals)
• Validate encryption, access controls, audit logging

Deliverables:

Monitoring dashboards, alerting system, security review doc, cost projection model

Phase 2: Orchestration & Decision Logic

(Weeks 7–12 ½-time | 4–8 weeks full-time)

Agent orchestration

• Design multi-agent workflow orchestration
• Implement job queues, polling, webhooks, retries, timeouts
• Build agent health monitoring and failover to manual workflows

Decision & escalation intelligence

• Auto-verify vs manual review logic
• Confidence scoring (data quality, source reliability, discrepancies)
• Explainable escalation routing

Performance & integrations

• Prompt optimization, caching, and cost controls
• API integrations (FSMB PDC, ABMS CertiFACTS, NPI, OIG/SAM, others)
• Structured/unstructured response parsing

Deliverables:

Orchestration framework, decision logic system, 3+ API integrations live, 20–30% cost-per-verification reduction

Phase 3: Scale & Production Readiness (Ongoing)

• Scale to 25K+ wallets and 150+ organizations
• Rate limiting, circuit breakers, graceful degradation
• Immutable audit trails with 7-year retention
• Agent performance analytics, anomaly detection
• Benchmarking and regression testing

​

What You’ll Bring

Required (5–8 years)

Production AI systems

• Shipping LLM-powered systems to production
• Building reliability on non-deterministic models
• Debugging hallucinations, regressions, agent failures
• AI observability and monitoring

Technical

• LLM systems: prompting, error handling, cost management
• Backend: TypeScript/Node.js, PostgreSQL, APIs, job queues
• Orchestration: LangChain/LlamaIndex-style frameworks
• Security: encryption, access controls, audit logging
• Observability: metrics, logging, alerting, distributed debugging

Strongly Preferred

• Healthcare compliance (PII, Joint Commission, NCQA)
• Web automation/scraping (Playwright, Puppeteer)
• Multi-agent coordination patterns
• LLM cost optimization

​

Work Style

• Comfortable in pre-revenue ambiguity
• Balances speed with compliance
• Strong async communication
• Effective with external contractors
• Pragmatic about technical debt

​

What You’ll Get

Impact

• Architect the intelligence layer of a first-to-market AI verification system
• Direct contribution to revenue milestones and scale

Learning

• Deep production LLM and agent orchestration experience
• Regulated healthcare systems engineering
• Fractional CTO-level strategic ownership

Flexibility

• Fully remote, async-first
• Deliverables over hours
• Flexible schedule with core overlap

Compensation

• $8K+/month (scales with hours + experience)
• Potential conversion to full-time CTO role
• 0.25%–6% equity (PT vs FT 1-year cliff)

​

How We Work

• Kanban, weekly delivery cycles
• Weekly demos (async-recorded)
• Daily async standups
• Weekly sync (45 min)
• Tools: GitHub, Slack
• Definition of Done: merged, deployed, documented/demoed
• Target: 5–7 medium stories/week (quality > quantity)

​

Interview Process

1. Intro call (30 min)
2. Technical deep-dive (60 min)
3. Async code/architecture review
4. Culture fit (30 min)
5. Paid trial (1 week) — ship monitoring or orchestration component

Timeline: 1–2 weeks end-to-end

​

How to Apply

Note: This is a syndicated job post. Fractional Jobs found it on the web, but we are not working with the client directly, so we don't have control over or knowledge of the application process. To apply, click on the "View Application" button and follow the application's instructions. Let us know how it goes!