Raman Marozau

CTO & Founder of Target Insight Function

Principal Engineering Architect

Skip to main content

Build

Systems That Hold

Platforms, protocols, and production tooling — grouped by product ownership, not engineering function.

0

Platforms Shipped

With dedicated domains

0

Production Packages

npm · PyPI

0

Protocols & Frameworks

ACP · MCP · SSR

0 / 0

Active / Total

Maintained

Available for architecture & platform consulting

Contact Raman
npm
PyPI
GitHub
Web
PROTOCOLS

stdio Bus Protocol & Tooling

stdiobus.com
Visit Website →

Deterministic C kernel, agent transport, ACP/MCP protocol bridges.

stdio Bus – Agentic Orchestration Ecosystem

GitHub

Deterministic C kernel providing transport-level routing for ACP/MCP-style agent protocols. The core runtime behind the stdio Bus ecosystem.

c
runtime
deterministic
Open
Active

Universal MCP Module Launcher

npm

One command to run any MCP server module — in any language, on any platform. A manifest-driven launcher for stdio Bus ecosystem that replaces hardcoded bash commands in mcp.json with a portable, runtime-agnostic mcpx run <module>.

stdiobus
mcp
model-context-protocol
Open
Active

MCP Agentic — Multi-Agent Orchestration Server

npm

Agent orchestration server that connects MCP clients to ACP-compatible agents through stdio Bus.

stdiobus
mcp-agentic
mcp
Open
Active

stdio Bus C++ SDK for AI Agent Transport

vcpkg

A C++17 SDK for stdio Bus — a process-embedded message bus for AI agent orchestration. Manages child worker processes communicating over stdin/stdout using JSON-RPC (MCP/ACP protocols).

stdiobus
mcp
acp
Open
Active
Specification

Normative protocol specification with RFC 2119 requirements.

Getting Started

Build instructions, configuration, and minimal examples.

Downloads

Source repository, versioning, and release information.

MCP Agentic Skills

npm

Structured, validated, machine-readable SKILL capabilities and expertise for AI agents — delivered over Model Context Protocol via stdio Bus

stdiobus
mcp-agentic
mcp
Open
Active

stdio Bus Rust SDK for AI Agent Transport

crates

An async Rust SDK for stdio Bus — a process-embedded message bus for AI agent orchestration. Manages child worker processes communicating over stdin/stdout using JSON-RPC (MCP/ACP protocols).

stdiobus
mcp
acp
Open
Active

stdio Bus – Workers Registry

npm

Worker implementations for stdio Bus kernel — ACP, MCP, and protocol bridges.

stdiobus
acp
mcp
Open
Active

stdio Bus Node.js SDK for AI Agent Transport

npm

Native Node.js binding for stdio_bus — AI agent transport layer. No external binary required.

stdiobus
node
native
Open
Active

Agent-to-agent ACP Chat

npm

ACP Chat — agent-to-agent dialogue over TCP/Unix socket using the ACP protocol.

acp
agent
chat
Open
Active

stdio Bus Python SDK for AI Agent Transport

PyPI

Python SDK for stdio_bus — async/sync API, Docker and native FFI backends, cross-platform AI agent transport layer.

python
sdk
stdiobus
Open
Active
FRAMEWORKS & DELIVERY

Runtime Web Platform

runtimeweb.com
Visit Website →

Unified TypeScript runtime platform for backend, infrastructure, and web delivery across local and AWS environments.

Runtime Web — Unified Serverless Framework for Full-Stack TypeScript on AWS

npm

TypeScript-first framework for building backend, infrastructure, and frontend in a single codebase — context-aware execution, type-safe DI, and native AWS integration.

runtime
ssr
lambda
Open
Active
Unified Codebase

Backend, infrastructure, and frontend in a single TypeScript project — one repo, one build, full type safety.

Context-Aware Execution

Same code adapts to web, Lambda, and local development contexts automatically.

Type-Safe DI

Compile-time dependency injection with automatic service resolution — no runtime wiring errors.

AWS-Native Integration

First-class APIs for Lambda, API Gateway, DynamoDB, and S3 with zero-config deployment.

Real-world @worktif/runtime handler — AI chat with memory and vector search on AWS.

npx @worktif/runtime init --template ai-chat

typescript
import type { LambdaDefinition, LambdaEvent } from '@worktif/runtime';
 
/** AI chat microservice with memory + vector search on AWS **/
export const chatHandler: LambdaDefinition<ChatTies> = {
  ties: { ai: Bedrock, memory: DynamoDB, search: OpenSearch },
 
  http: { method: 'POST', path: '/chat' },
 
  handle: async ({ body, ties }: LambdaEvent<ChatTies>) => {
    const history = await ties.memory.get(body.sessionId);
    const relevant = await ties.search.knn(body.message, 5);
 
    const reply = await ties.ai.invoke({ message: body.message, history, relevant });
 
    await ties.memory.append(body.sessionId, { user: body.message, assistant: reply });
    return reply;
  },
};
Show 15 more lines
AWS MARKETPLACE

Lambda Kata AWS Marketplace

lambdakata.com
Visit Website →

AWS Lambda runtime layer, CDK integration, and tamper-resistant licensing.

Lambda Kata

cloud

Lambda Kata is a high-performance runtime that removes cold starts and stabilizes p95, p99 latency for Node.js Lambda functions

aws
cdk
lambda
Open
Active
Predictable Latency

Stable response times by minimizing cold-start variance in Node.js Lambda workloads.

Runtime Layer

Runs as a native Lambda runtime layer, improving execution consistency without app rewrites.

CDK Integration

Integrates directly with AWS CDK for repeatable setup, versioning, and deployment workflows.

No Warmup Hacks

Upgrades the runtime under existing Node.js Lambdas — same APIs, domains, and infrastructure.

Lambda Kata for AWS CDK integration

cloud

Built specifically for AWS Lambda — operates entirely within your AWS accounts as a runtime layer for Node.js functions.

aws
cdk
lambda
Open
Active

Lambda Kata Licensing

npm

Tamper-resistant native licensing validator for Lambda Kata Integration.

licensing
security
lambda
Open
Active
CONSULTING & INFRASTRUCTURE

Cloud & Infrastructure

worktif.com
Visit Website →

AWS infrastructure toolkits and the Target Insight Function platform.

Target Insight Function

Web

Full-stack engineering platform — the foundation for AI infrastructure, quantum algorithm research, and real-time system performance.

platform
ai
infrastructure
Open
Active
Critical Scope

Each engagement targets clearly defined high-impact systems — stabilize what matters most before scaling further.

Concrete Delivery

Pull requests, infra changes, and technical notes — not abstract recommendations.

Fragile Flows

Release-critical Lambda logic accumulates edge cases and becomes too risky to modify.

Release Confidence

Standardized runtime and infrastructure patterns reduce deployment risk and improve day-to-day operability.

SCIENCE

Research & Governance

AI data governance frameworks and licensing innovation.

Theory of Everything Decoherence Validation: Empirical Validation of the ToE Consistency Relation

GitHub

Foundational research repository for decoherence-driven validation work toward a unified Theory of Everything hypothesis.

physics
theory-of-everything
decoherence
Open
Research

Internal Decoherence in a Closed-System Cosmology Induces an Effective Stress-Energy Component and Governs Early-Universe Dynamics. Modeling

GitHub

We present a technical validation for complete, testable cosmological framework in which the Universe begins in a globally coherent quantum state and the first local act of decoherence inside the closed system creates its own environment. This repository contains a Python-based research module designed to explore quantum entanglement and decoherence as drivers for the expansion and materialization of field modes in closed quantum systems. The modeling focuses on a theoretical cosmological model where the stress-energy tensor emerges dynamically from decoherence processes.

cosmology
physics
python
Open
Research

Purem – advanced high-performance computational library

PyPI

High-performance computational library for vectorized mathematical operations — up to 429× faster softmax, exp, and element-wise transforms via optimized assembly on x86-64, ARM64, and Apple Silicon.

ai
compute
acceleration
Open
Active
PyPI v3.0.6

Experimental Validation of the L-Function for Task Optimization in Multi-Agent LLM Systems

GitHub

Experimental validation of the L-function, a novel mathematical construct designed to optimize task prioritization, contextual coherence, and resource management within Large Language Model (LLM)-driven Multi-Agent Systems

llm
multi-agent
validation
Open
Experimental

Open Data Ownership License (ODOL) 18.0

GitHub

Empowering Data Ownership in the AI Domain — a license framework for AI data governance.

license
ai
data-ownership
Open
Active
LABS & R&D

Labs / R&D Forks

Research forks, upstream contributions, and experimental branches.

Bitcoin Core integration/staging tree → Experimental distributed R&D workflow

GitHub

Experimental Bitcoin R&D focused on integrating stdio Bus protocol patterns for deterministic distributed workflow orchestration.

bitcoin
blockchain
distributed
Open
R&D Contribution

AI agents running research on single-GPU

GitHub

AI agents running research on single-GPU nanochat training automatically — integrated with stdio Bus protocol. Fork of Karpathy autoresearch.

ai
research
gpu
Open
Research

Codex is OpenAI’s coding agent for software development

GitHub

Terminal coding-agent R&D fork with active contributions around stdio Bus interoperability and agent workflow experiments.

ai
coding-agent
terminal
Open
R&D Contribution

ACP Registry

GitHub

Registry and reference surface for ACP ecosystem coordination, compatibility, and implementation discovery.

acp
agents
registry
Open
R&D Contribution

Agent Client Protocol

GitHub

Protocol workstream for agent-client interoperability, transport contracts, and cross-runtime integration patterns.

acp
protocol
editor
Open
R&D Contribution

Kiro is an agentic IDE

GitHub

Active fork contribution stream with bug reports, feature proposals, and implementation feedback for agentic IDE workflows.

ide
agentic
ai
Open
R&D Contribution
DEVELOPER TOOLING

Developer Tooling

CLI automation, versioning, prompt engineering, and AI chat templates.

TypeScript utilities library

npm

TypeScript utilities library providing dependency injection, logging, exception handling, and Zod validation tools with AWS Lambda integration.

typescript
di
logging
Open
Active

Unix shell automation library

npm

TypeScript-based Unix shell automation library for command-line operations, scripting, and terminal management with comprehensive CLI integration.

unix
shell
automation
Open
Active

AWS infrastructure toolkit

npm

TypeScript-based AWS infrastructure toolkit: DynamoDB integration, AppSync support, SES functionality, GraphQL capabilities with audit logging and AWS Signature V4 authentication.

aws
dynamodb
graphql
Open
Active

Production-ready UI utility package for React 19 + MUI 7

npm

Material UI Theme Provider and Customization Suite for React applications with dark mode support and dynamic color schemes.

react
mui
theme
Open
Active

Versionings — Semantic Release Platform for Git

Web

CLI tool that automates semantic versioning workflows for Git repositories. Bumps versions, creates branches and tags, and optionally pushes changes and opens pull requests on supported SCM platforms.

versioning
semver
automation
Open
Active

Chat Application Starter Kit

GitHub

AI chat application template — a starter for building agent-powered conversational interfaces.

ai
chat
template
Open
Experimental

Prompt Engineering Repository

GitHub

Repository for basic and extensive prompt engineering including prompts, data, descriptions, specifications, and more.

prompt-engineering
ai
specifications
Open
Active
Target Insight Function Logo

Cloud Architect & Systems Innovator. One Vision. Infinite Scale.

HomeBuildResearchComposeExperienceEducationContact
raman@worktif.com+48 881 592 968

Szwedzka 4, Warszawa

© 2026 Target Insight Function Ltd. All rights reserved.