Chadsel LLC
Institutional AI.
Your Data Never Leaves.
By Chadsel Chen, Managing Principal ·
We deploy AI agents for hedge funds, RIAs, and family offices with one constraint: trade data, MNPI, and portfolio positions never transit third-party APIs. On-premise LLM inference, isolated Azure or AWS tenants, and SEC/FCA-compliant architectures.
OpenClaw, Ollama, LangChain, and LlamaIndex — the institutional AI stack configured for your compliance requirements and data classification.
Deployment Architecture
Three models. One requirement: MNPI stays yours.
Every institutional AI deployment starts with data classification. We recommend the architecture after understanding what data your workflows touch — not before.
On-Premise
- Data location
- Your infrastructure
- MNPI exposure
- Safe
- Compliance posture
- Maximum
- Best for
- MNPI, trade data
Isolated Cloud
- Data location
- Dedicated Azure tenant
- MNPI exposure
- Configurable
- Compliance posture
- Strong
- Best for
- LP reporting, research
Hybrid
- Data location
- Split by classification
- MNPI exposure
- Routed on-premise
- Compliance posture
- Configurable
- Best for
- Most institutions
Agent Stack
The platforms we deploy.
Purpose-selected for institutional workflows, data security requirements, and the compliance constraints of investment management.
OpenClaw
Open-Source IntegrationOpenClaw agent integration for capital markets. Connects to Bloomberg, FactSet, OMS/PMS, and prime broker data — deployable on-premise or in an isolated Azure or AWS tenant. No trade data leaves your infrastructure. Purpose-built for the regulatory complexity and performance standards of institutional investment management.
Claude (Anthropic)
Reasoning model for LP report drafting, compliance document analysis, and regulatory Q&A. Deployed in an isolated Azure or AWS tenant — no data shared across tenants.
Codex / Computer-Use
Automation agents for reconciliation workflows, data normalization, and repetitive back-office operations across Bloomberg, prime broker portals, and internal systems.
Open Harness
Agent orchestration with data classification routing — MNPI-flagged data stays on-premise; general research queries route to cloud models with audit trail.
Ollama
On-premise LLM inference. Runs Llama, Mistral, and Phi models on your hardware. Zero egress risk for MNPI, position data, or anything subject to Regulation S-P.
LangChain
Orchestration framework connecting AI models to Bloomberg feeds, internal research databases, compliance systems, and custodian APIs.
LlamaIndex
Indexes internal research, DDQ responses, LP documents, and regulatory filings for semantic search and retrieval-augmented generation — fully on-premise.
Investment
Institutional AI deployment pricing.
Engagements scope to your compliance requirements and data classification. Most institutions start with a focused pilot before expanding to full-stack deployment.
| Engagement | Starting From | Typical Range |
|---|---|---|
| AI readiness & governance assessment | $5,000 | $5,000–$12,000 |
| Private AI / agent pilot | $20,000 | $20,000–$60,000 |
| Institutional production deployment | $75,000 | $75,000–$250,000+ |
| Specialist consulting | $175/hr | $175–$275/hr by labor category |
| Managed AI operations | $2,000/mo | $2,000–$6,000/mo |
Ranges exclude third-party licenses, hardware, cloud consumption, model usage, and compliance testing unless listed in the signed scope.
Infrastructure Stack
Institutional AI infrastructure.
Air-gapped, auditable, and designed for the data security requirements of investment management.
| Layer | Technology |
|---|---|
| Local LLM Runtime | Ollama, vLLM (air-gapped) |
| API Proxy | LiteLLM (OpenAI-compatible) |
| Agent Orchestration | LangChain, Open Harness |
| Document Intelligence | LlamaIndex, pgvector, Qdrant |
| Institutional BI | OpenClaw (Bloomberg, FactSet, OMS) |
| Cloud AI Isolated | Azure AI Foundry (dedicated tenant), AWS Bedrock |
| Containerization | Docker, Kubernetes, KServe |
| Data Security | Encryption at rest/transit, RBAC, immutable audit trail |
| Open Models | Llama 3, Mistral, Phi-3 (on-prem); GPT-4o, Claude (isolated cloud) |
Compliance Architecture
Built for auditors, not just operators.
Every AI deployment we build for financial services clients includes the controls an examiner would look for: audit trails, access logs, data residency documentation, and incident response procedures.
We have built and operated cybersecurity programs that passed both SEC and FCA examinations. That background shapes how we design AI architectures — defensively, with documentation that holds up under scrutiny.
Regulation S-P
Data handling and access controls reviewed against SEC Regulation S-P requirements before deployment.
Rule 17a-4
Immutable audit log architecture for AI-generated outputs subject to books-and-records obligations.
NIST SP 800-53
Zero-trust access, encryption at rest and in transit, and incident response aligned to NIST 800-53.
FCA SYSC
Operational resilience and data governance architecture reviewed against FCA SYSC 8 requirements for UK-regulated entities.
Common Questions
Frequently asked.
How do you ensure MNPI never reaches external AI providers?
For MNPI-sensitive workflows, we deploy local inference — Ollama or LlamaIndex on your own hardware, or an air-gapped Azure Stack instance you control. No trade data, portfolio positions, or material non-public information transits third-party APIs. The architecture is reviewed against SEC Regulation S-P and FCA SYSC data handling requirements before go-live.
Are your AI architectures SEC and FCA compliant?
Our deployments are designed around Regulation S-P, Rule 17a-4, NIST SP 800-53, and FCA SYSC requirements. Full audit trails, access controls, encryption at rest and in transit, and data residency documentation are included in every institutional engagement. We have managed IT and cybersecurity programs that passed SEC and FCA audits.
What deployment models are available for institutional clients?
We offer on-premise (bare metal or Azure Stack, air-gapped option), cloud-isolated (dedicated Azure Government or commercial tenant with no cross-tenant data sharing), and hybrid (MNPI stays on-premise, general research uses cloud). We assess your compliance requirements and data classification in week one before recommending an architecture.
Can AI agents handle MNPI (material non-public information) safely?
Yes, with the right architecture. MNPI requires a fully air-gapped local LLM — no data should touch any external inference endpoint. We deploy Ollama or vLLM on dedicated hardware with RBAC, full audit logging, and encrypted storage. The architecture is documentable for SEC or FCA examination on request.
What compliance frameworks do your financial services AI deployments follow?
We design deployments to align with SEC Regulation S-P (client data privacy), Rule 17a-4 (electronic record-keeping), NIST SP 800-53 (information security controls), and FCA SYSC (systems and controls). For firms subject to both SEC and FCA, we can design a dual-jurisdiction architecture. All deployments include a written data flow diagram suitable for regulatory examination.
New to Chadsel LLC?
Start with AI readiness and governance.
A $5,000–$12,000 assessment defines MNPI boundaries, governance controls, target workflows, deployment architecture, and pilot acceptance criteria.
(212) 749-9408