Senior AI Architect – Multi-Agent Systems & Platform Infrastructure

Senior AI Architect – Multi-Agent Systems & Platform Infrastructure

Location:

Remote (US Preferred)

Compensation:
Equity-first (Pre-Seed) + Competitive Salary Post-Seed

Equity Range: 2.0% – 3.0%

Reports To:

Director of AI Systems & Orchestration / Head of Engineering

Seniority:
Senior-Level (Hands-On Architect)

Time Commitment:
Fractional (8–12 hours/week, pre-seed equity-based)

About Nivalto + AURA

Nivalto is building AURA — the world’s first fiduciary-grade agentic operating system for ultra-high-net-worth families and their advisors. AURA orchestrates trust management, gifting, estate workflows, private investments, and family governance through a secure, multi-agent architecture grounded in privacy, compliance, and long-term legacy planning.

At its heart is an advanced orchestration layer where intelligent agents coordinate across documents, assets, decisions, advisors, and heirs — bringing real-time intelligence and advisor consent into wealth management for the first time.

Nivalto was recently profiled in Modus as “the stealth-mode startup building an army of agents for family offices,” recognizing its novel approach to AI-native orchestration, simulation logic, and multi-role collaboration.

Why Now?

Recent 2024–2025 Family Office Reports reinforce the urgency and scale of the market Nivalto is addressing:

• Family offices are growing rapidly: from 8,030 in 2024 → 9,030 in 2025 → 10,720 by 2030
• AUM is projected to grow from $3.1T to $5.4T by 2030
• 89% feel under-invested in technology, and 49% are launching digitization initiatives
• Only 11% currently use AI, but 94% expect to adopt AI in the next 5 years
• Yet, fewer than 15% use AI for automation or compliance tasks — a gap Nivalto is built to close

Your Role:
Senior AI Architect, Multi-Agent Systems & Platform Infrastructure

As the lead AI architect at Nivalto, you will build the orchestration and intelligence layer powering the world’s first fiduciary-grade multi-agent platform. You’ll architect memory, consent, and execution flows — and directly code the infrastructure that empowers advisors, families, and intelligent agents to collaborate securely and scalably.

Key Responsibilities

• Architect and implement AURA’s multi-agent orchestration framework, including agent routing, vector memory, and bringing real-time intelligence and secure, role-aware coordination into wealth workflows.
• Design prompt engineering pipelines, retrieval-augmented generation (RAG) systems, and persona-aware memory recall logic.
• Develop secure APIs for coordinating agent outputs with user roles, documents, and structured data sources.
• Design and implement distributed observability and tracing frameworks (e.g., Open Telemetry, Jaeger) to monitor agent workflows, performance, and compliance in real time.
• Create a reusable library of secure, auditable agent actions — enabling modular and deterministic orchestration workflows.
• Design and optimize context ingestion pipelines across Vault, Circle, and external APIs to enable rich, persona-aware agent intelligence.
• Embed privacy-aware, regulatory-aligned agent workflows (IRS, SEC, FINRA, HIPAA) into the orchestration layer.
• Collaborate on SOC2-ready audit frameworks, consent middleware, and redaction systems.
• Design safeguards and overrides to ensure human-in-the-loop (HITL) control where needed.
• Oversee the integration of consent and compliance controls, ensuring agent actions adhere to user permissions and regulatory standards, with comprehensive auditability.
• Implement zero-access encryption, role-based access control (RBAC), and privacy validation for all agent data flows and logs.

Hands-On Engineering Leadership

• Write production-grade code and prototypes to drive PoC delivery, agent development, and orchestration testing.
• Collaborate with backend and ML engineers to integrate OpenAI, Anthropic, Pinecone, Lang Chain, and Semantic Kernel components.
• Develop and refine test plans, evaluation pipelines, and debug tools for cross-agent messaging and decision flows.
• Lead evaluation-driven development (EDD) by defining live…