Founder · AI Product Engineer · Zirvox

Builder.
Not just
a developer.

I turn ideas into real, usable products fast.

AI systems, dev tools, and data-driven products with real commercial impact.

View ProductsOpen to Work
ABOUT

Self-taught.
Execution-driven.

I don't stay in the concept phase. I define the problem, build the product, and ship it in short cycles. I build technical tools with real commercial potential: dev tools, AI infra, and data systems that solve actual problems.

Ship first. Refine later.

STACK

Technology I
actually use.

EXPERIENCE

5+ years.
Real systems.

2020 · Present

AI Product Engineer

Founder @ Zirvox · Independent

Designed and shipped multiple production systems. Built LLM-powered products: agents, document processors, and prompt pipelines running in production environments with real users.

Real Impact
▸Reduced p95 latency from 1.8s to 680ms avg response time (AI pipelines) via query batching and Redis cache
▸Cut LLM costs 40% with model routing: agile models for speed, robust models for accuracy
▸Eliminated 60% of manual workflows through AI pipeline automation
▸99.5%+ uptime across all production environments
PROOF

Built things that run.

Products with real users, real infra, and real constraints.

Latency
1.8s p95680ms avg response time (AI pipelines)

Query batching + Redis cache

LLM Cost
GPT-4 only40% less

Model routing: agile models for speed, robust models for accuracy

Manual Work
Manual ops60% cut

AI pipeline automation across workflows

Products
Side projectsUsed in production systems

Active projects, heavily adopted by open-source communities

PRODUCTS

What I built
that actually works.

Founder @ Zirvox · two products in development. Four OSS published.

ZIRVOX
Zirvox AI
AI · In Development
ZIRVOX · AI PRODUCT · COMING SOON

Zirvox AI

Zirvox an ecosystem of AI products, dev tools, and data platforms. AI platform built from scratch, with proprietary infrastructure, without relying on third-party SDKs.

▸Proprietary AI infrastructure
▸Real production pipelines
▸Average latency below 200ms
Custom AI Infra · FastAPI · Next.jsComing Soon
ZIRVOX
FlipperZ
ZIRVOX · RESEARCH AGENT · COMING SOON

FlipperZ

Research-to-deliverable agent. Transforms a query into a professional document with statement-level verifiable provenance.

Next.js · TypeScriptComing Soon
LicensePulse
DEV TOOLS · OSS INTELLIGENCE · OPEN SOURCE

LicensePulse

OSS license watchdog. Monitors the dependencies you use and alerts you when a license changes, before it becomes a legal problem.

Node.js · DockerView repo
LeadStrike
DATA · REVENUE INTELLIGENCE · OPEN SOURCE

LeadStrike

B2B prospecting engine by reputation gap. Finds companies losing money right now and delivers the lead ready to close.

Node.jsView repo
Promptinel
AI INFRA · LLM OBSERVABILITY · OPEN SOURCE

Promptinel

Drift monitor for LLM pipelines. Detects when your prompts' behavior changes silently, before the user notices.

TypeScript · Node.jsView repo
PublishCanary
DEV TOOLS · NPM SECURITY · OPEN SOURCE

PublishCanary

Scans the actual npm tarball before publish. Blocks source maps, secrets, and internal files before they leak to the public.

Node.js · npmView repo
AI ENGINEERING

LLMs are not buzz.
They are infra.

RAG pipelines, multi-model routing, memory layers. Not as demos, but as production systems that deal with real latency and token limit constraints.
Less wrapper. More engineering.

inference_pipeline.rs

Conversational Agents

Proprietary · Open Weights

Document Processing

Automated extraction + structuring

RAG Pipelines

Embeddings + vector search

Prompt Engineering

Structured pipelines

On-Device Inference

TFLite · ONNX -70% Cost

LLM Orchestration

Multi-step LangChain

Zirvox an ecosystem of AI products, dev tools, and data platforms.

Looking to build
real AI products?

If it needs to ship, I can build it.Let's build
GitHubLinkedIn
Available for projects · 2026