What makes this different from staff augmentation?

One person with seven specializations unavailable in staff aug. COBOL PICTURE-clause forensic find in 127 minutes that 3 years of external audits missed. 3-5x lower cost for broader coverage.

Is the sovereign inference cluster operational?

Yes. 55 commits ahead of upstream on a production fork of exo (upstream lineage). RDMA ring topology. 389ms P50. AWS Bedrock and Azure OpenAI format-compatible. $26,400/mo cloud inference eliminated.

What audit-trail artifacts does the sovereign inference build produce?

SHA-256 hash-chained chain-of-custody on every inference output. Quality-management-system documentation, bias logs, and human-oversight records with a deterministic event-sourced audit trail, operational now.

What compensation are you targeting?

Negotiated privately, by role and scope — remote · W2 · skills-first.

100% remote. US-based, Mountain Time. Async-first. No relocation.

What kind of engineer are you?

Agentic architect and AI/data engineer — agentic systems, sovereign inference infrastructure, and adversarial-AI validation, every result replay-verifiable.

What proof do you offer?

Deterministic replay artifacts. SHA-256 hash-chained chain-of-custody. COBOL PICTURE-clause forensic audit in 127 minutes. 27× throughput. 96% latency cut. 0 false negatives across the public adversarial corpus (140/140 caught, n=220). All verifiable at puodziukas.dev/proof.

Why hire you instead of engaging a consulting firm?

Single operator, full-time, inside your perimeter. The person who scoped the work does the work — every output logged and replay-verifiable.

Yes. Standard MNDA on day one. All work product stays air-gapped.

What stacks do you work in?

COBOL · Rust · Python · TypeScript · Next.js · LLM APIs · AWS · GCP · Terraform · Pulumi · OpenTelemetry · NIST FIPS 203/204/205.

puodziukas.dev · remote · available Q3 2026

Multi-Agent Systems Architect

Multi-agent council. SHA-256 hash-chain attestation. continuous internal-harness invocations. Zero message loss.

8 specialized agents (1 orchestrator + 7 specialists). continuous internal-harness invocations. SHA-256 hash-chained. Every step reproducible.

reproducible · verified · available Q3 2026

The problem

A representative failure mode in multi-agent handoffs: Agent B passes a truncated context window to Agent C, and Agent C returns a confident wrong answer. The framework does not catch it, and the audit log records success.

What exists, reproducible

The multi-agent autonomous council runs continuous internal-harness invocations with SHA-256 hash-chained attestation on every step. Handoff failures are caught at the boundary, not post-hoc. Every output is replay-verifiable.

Relevant projects

→Multi-agent autonomous council: SHA-256 hash-chain attestation, continuous internal-harness invocations, zero divergence
→OODA-loop state machine: task delegation, context preservation, escalation logic
→Human-in-the-loop failsafe with signed audit trail and deterministic replay of any execution path
→LangGraph-class orchestration: reference architecture across 6 regulated domains
→Context fidelity proof: deterministic replay over a 47,312-record deterministic replay corpus, zero coherence degradation
→Game-theory routing: confidence threshold → model selection (no token waste, no wrong model)
→Merkle-chain audit log with every agent action hash-linked, tamper-evident, exportable

What you receive in 30 days

State machine design doc: node graph, escalation paths, memory spec (Week 1)
Single-agent POC with SHA-256 hash-chain + audit log running in your environment
Human-in-the-loop failsafe tested with 100 adversarial edge cases

Companies at this intersection

LangChain · CrewAI · Microsoft AutoGen · Cognition · Letta · Adept · MultiOn · Lindy

Working terms

W2 remote · async-first · deliverable-gated output · first artifact in 48h · available Q3 2026

Work with me →

WRITTEN SCOPE FIRST · MILESTONE-GATED · REPLAY-VERIFIED

Full hire page →

All results under NDA. Representative figures. SHA-256 hash-chained audit trail available on request. · puodziukas.dev