Simas Razinskas

Proof

Proof, in one place

Everything a team evaluating Simas Razinskas needs on one page: twelve case studies, public code, concrete service scopes and a one-page brief you can forward.

Deciding whether to book a call usually comes down to one question: has this person built the kind of system we need? This page is the shortest path to that answer — case studies, public code, and a brief you can forward to whoever else is deciding.

12

production case studies

AI systems, job queues, deployment infrastructure, media pipelines, SaaS isolation and automation.

12+

years writing code

Senior engineering judgement across backend, frontend, infrastructure and applied AI.

EN/LT

bilingual delivery

Remote or in Vilnius, with working fluency in both English and Lithuanian contexts.

GitHub

engineering in the open

Public code and commit history back the anonymized private-system case studies.

Representative case studies

Client names stay private, so each system is described by what it does: the constraint, the design, and how it behaves when things fail.

  • Zero-downtime blue-green deploys with health-gated rollback

    Only a tested image tag can be promoted, traffic moves after health passes, and a failed release rolls back before users see it.

    Docker Compose · Traefik · Postgres · shell orchestration
  • An agent-operated task engine

    Agents can claim, heartbeat, complete, fail, and escalate work safely while humans watch the same board update in real time.

    Drizzle · Postgres LISTEN/NOTIFY · SSE · contract-first API
  • Idempotent webhook ingestion at scale

    Data-loss risk moved from a quarterly investigation to continuous checks that detect gaps, repair them, and expose operational health.

    Go · Postgres (SKIP LOCKED, matviews) · HMAC · advisory locks
  • Frame-accurate video localization

    Long clips stay locked to source timestamps while the original non-vocal audio survives beneath the synthesized dub.

    WhisperX · XTTS voice cloning · Demucs · FFmpeg · CTranslate2

What the work looks like

Four concrete engagements: diagnose the workflow, harden the prototype, design the agent operating model, or steady the system already in motion.

Where the work lands

The strongest fit is where AI behavior, backend reliability, internal operations and deployment constraints meet.

  • LLM media-localization pipelines

    End-to-end translation and adaptation of video, audio and image content — transcription, machine translation, voice synthesis and on-image text replacement, wired together to run at scale.

  • Multi-service automation platforms

    Monorepos stitching dashboards, queues, APIs and third-party integrations into the daily operational workflow of marketing and ops teams.

  • Agentic AI tooling & assistants

    Retrieval-augmented helpdesk assistants, automated ticket triage and internal copilots — LLMs put to work against real business data and processes.

  • Browser & design-tool plugins

    Figma plugins and Adobe scripting for design automation, plus Chrome extensions that fold internal tooling directly into the browser.

A brief you can forward

One generated PDF page — fit, case studies, engagement shape and the booking path — for whoever else needs convincing.

Download the one-page brief

The public record

Code, commit history and a decade of shipped work — verifiable without taking my word for it.

Proof

  • Case studies before claims — every service links to a system that shipped.

  • The engineering is public even where the clients can't be.

  • One page per service, so scope is visible before the first call.

  • A one-page brief you can forward, generated from this site's content.

Have a system that has to work?

Book a call and we'll map it out: what it should do, where it will break, and the smallest version worth building first.