RZMRN
EN/PL

Maksym (Max) Beiev — Applied AI Engineer · AI & Automation Lead

AI agents & automation / Model fine-tuning & inference / Evaluation & quality / 10+ years in video production

Explore My WorkLet’s Talkgithub.com/Beiev↗
$cat stats.json
>years_in_production: 10+
>client_projects_delivered: 298
>ai_assets_in_first_4_months: 4,000+
>working_languages: 4
$chronicle tail -n 6
2026-06-04 20:51gitchronicle: resilient WAL init under concurrent cold-start
2026-06-04 21:13gitchronicle: decommission dead agent jobs — honest > impressive
2026-06-04 22:24gitchronicle: native hybrid recall — FTS5 + vector + temporal
2026-06-04 22:44milestonechronicle v9 shipped: crash-safe I/O, 185 tests green, 5 launchd jobs
2026-06-07 13:48milestone[client·nda] LoRA 4-arm ablation locked: DoRA r32 wins — r64 went plastic
2026-06-10 00:15gitportfolio: new GenAI pipeline case study + honesty pass on metrics

# real events from my open-source memory system — sanitized, timestamps true

Scroll
Open to: Applied AI Engineer · AI & Automation LeadFull-time / Long-term B2BFully remote · International teamsLocation-flexible B2B · European time-zone overlapEN · PL · UA · RU

Selected Work

Content Production
AI Systems

Video · Motion · Live · Post-Production

Automation · Multi-Agent · Programmatic Video · Pipeline

Production experience. Technical delivery. A team that knows how to use what we build.

The AI studio I built — and the agents that run it
SYSTEMS

The AI studio I built — and the agents that run it

≈6× Lower compute cost per video44% → 81% Output run by agents, in one month89% → 0% Frames with the wrong face

From the studio’s own reports and logs, anonymized.

ComfyUIOpen-weights image & video modelsLoRA / DoRAPythonFastAPI
View Case Study →
Head of Content Production at an e-learning company — 4.5× faster pipeline
SYSTEMS + PRODUCTION

Head of Content Production at an e-learning company — 4.5× faster pipeline

4.5× Faster finishing~600 Lectures produced20 Courses delivered
Premiere ProAfter EffectsExtendScriptMidjourneyRunway
View Case Study →
298 Fiverr Orders — 120 US Clients — One Person
PRODUCTION

298 Fiverr Orders — 120 US Clients — One Person

298 Completed projects120 US clientsTop Rated Fiverr seller level
Premiere ProAfter EffectsDaVinci ResolvePhotoshopCinema 4D
View Case Study →
After Effects × Reels — Years of Craft in Seconds
PRODUCTION

After Effects × Reels — Years of Craft in Seconds

After EffectsPremiere ProCinema 4DPhotoshopMotion Design
View Case Study →
Director & lead video producer — Poland's national folk ensemble “Mazowsze”
PRODUCTION

Director & lead video producer — Poland's national folk ensemble “Mazowsze”

200+ Concerts filmed170+ PerformersMulti-cam Live production
Multi-camLive SwitchingDaVinci ResolveAfter EffectsDrone
View Case Study →
Parametric Motion — Where Code Replaces the Timeline
SYSTEMS + PRODUCTION

Parametric Motion — Where Code Replaces the Timeline

0 Manual keyframesReact Components as templates
RemotionReactTypeScriptNode.jsFFmpeg
View Case Study →
Max Chronicle — memory for AI agents, so they stop asking me the same question twice
SYSTEMS

Max Chronicle — memory for AI agents, so they stop asking me the same question twice

700+ Tests in CIRAG Hybrid retrieval: full-text, vector, timeMCP Server for any agent
PythonMCPSQLite FTS5Local embeddings (Ollama)Evals
View Case Study →
221 Lectures Rescued in 5 Automated Phases
SYSTEMS

221 Lectures Rescued in 5 Automated Phases

221 Lectures recovered600+ Lectures audited5 Automation phases
PythonBashGoodSyncOneDriveFFprobe
View Case Study →
RZMRN Digest — a 15-stage news pipeline that ran unattended for months
SYSTEMS + PRODUCTION

RZMRN Digest — a 15-stage news pipeline that ran unattended for months

15 Pipeline stages84 Sources2× Runs a day
Next.js 16React 19Tailwind v4Claude CodeMiniMax M2.5
View Case Study →
ls -la ./projects →

10+

Years in Production

298

Client projects delivered

4,000+

AI images & videos, studio’s first 4 months

4

Working Languages

I build the tools, bring the team together, and own the result.

My background is in production leadership. Today I use that experience to build AI workflows that people can rely on: generation pipelines, agents, custom model training, and quality checks that fit the work.

I built an AI content operation from the ground up and led its move from R&D into regular client delivery. That means shaping the product, setting priorities, developing the tools, supporting the team, and staying close to clients and the owners. The operation has delivered thousands of images and videos.

I’m looking for a fully remote role with an international team, where I can keep building useful AI systems and take responsibility for how they perform.

Need someone to take AI from idea to daily use?

I bring hands-on implementation, production experience, and team leadership. Open to fully remote Applied AI Engineer and AI & Automation Lead roles.

Let’s Talk