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

Solution / Applied AI Engineer

POC to reference implementation in weeks. 6 regulated domains. COBOL forensic proof closes enterprise deals.

The COBOL PICTURE-clause forensic record has a replay command. The board runs it live. That closes the deal.

reproducible · verified · available Q3 2026

The problem

A common enterprise-buyer objection: "We ran an LLM pilot last year, it hallucinated on our data, and six weeks of prompt engineering still didn't close the edge cases." What moves a skeptical prospect is a production system with a verifiable proof record, not another pitch deck.

What exists, reproducible

The COBOL PICTURE-clause forensic record ships with a replay command. The board can run it themselves, and that is what tends to close risk-averse buyers.

Relevant projects

→COBOL PICTURE-clause truncation found: proof reproducible with single command (board-meeting ready)
→the public 47,312-record COBOL probe, the POC that became the reproducible validation framework
→Domain-specific LLM wrapper for regulated financial environment: zero hallucinations in scope
→6 enterprise deployments: healthcare, finance, insurance, defense, mainframe, agentic
→ROI model per project: specific $ impact, not generic TCO claims
→First-artifact-in-48h standard on forensic scope: replay command ships before the debrief
→Domain-adjacent build experience: healthcare, banking/FINRA, insurance data environments

What you receive in 30 days

Technical discovery doc: client pain mapped to AI proof-of-concept (Week 1)
Working POC integrated with client primary data source
ROI model: $ impact quantified for 3 identified use cases

Companies at this intersection

Salesforce · ServiceNow · Workday · Zendesk · HubSpot · Brex · Rippling · Amplitude

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