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

AI Integration Engineer

COBOL → Rust → LLM. Reproducible forensics on a public corpus — not slideware.

The hard part of legacy integration is proving the arithmetic is correct before you move it. That part is public and clone-and-run.

reproducible · verified · available Q3 2026

The problem

The AI platform is ready, but the enterprise client's data lives in legacy batch outputs, DB2 tables, and COBOL programs from 1987. "Just use the API" runs into the fact that the API does not speak EBCDIC, and the batch arithmetic has not been independently verified.

What exists, reproducible

Start where the risk is: prove the legacy arithmetic before you move it. The COBOL PICTURE-clause forensic audit is public and clone-and-run — 47,312 records, $40,812.81 of silent truncation reproduced, 31 passing assertions. The COBOL→Rust migration ships byte-verified parity against the legacy output. That same deterministic-replay discipline is what maps batch outputs into LLM-ready context — built per environment, not a pre-canned connector.

Relevant projects

→COBOL PICTURE-clause arithmetic forensics: public probe, 31 pytest assertions, $40,812.81 reproduced across 47,312 records (clone-and-run)
→COBOL → Rust migration: byte-verified parity against the legacy batch output
→Deterministic-replay corpus: every transformation traceable to source record, SHA-256 hash-chained chain-of-custody
→COBOL static risk scanner: PIC truncation, ON SIZE ERROR gaps, EXEC SQL without SQLCODE — working regex rules
→Audit log: every transformation traceable to source record, hash-chained by design

What you receive in 30 days

Legacy arithmetic forensic audit: which programs, how much silent exposure (Week 1, replay command ships first)
First batch source mapped end-to-end into LLM-ready context, traceable to source record
Error handling + audit log with every transformation replay-verifiable

Companies at this intersection

Zapier · Workato · Boomi · Make · Apify · n8n · Pipedream · Merge API

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