CTO. Engineer. Builder.

AI products.
Built for the
real world.

I’m Kostyantyn. I turn complex workflows into useful software, from the first architecture decision to the team that ships it.

Explore my work Connect on LinkedIn

Currently building call2action as CTO

Kostyantyn Pshenychnyy
Hands-on engineering.
End-to-end ownership.
From ideaTo production

Building software since 2012

Across recruiting, legal tech & healthcare

Open-source author & published writer

How I work

The product and the system have to work together.

I work at the intersection of engineering leadership and hands-on delivery. That means defining what to build, making the architecture hold up, and helping a team get it into customers’ hands.

My recent work spans voice agents for recruiting, confidential document workflows and personalized healthcare. The common thread: make AI useful inside an actual business process, with clear boundaries and measurable outcomes.

More about my experience

Selected work

Complex problems.
Working products.

Three examples of the systems I’ve built and the decisions behind them.

call2action

CTO · 2025–present

A faster first conversation.

An AI-native recruiting platform connecting truck drivers with carriers. I own the technology strategy, architecture and engineering.

Read the engineering story
The problem
Recruiting leads lose momentum while they wait for a callback.
What I built
A driver marketplace connected to a carrier CRM, with AI voice outreach, structured applicant routing and call classification.
The outcome
Designed for AI-initiated contact within approximately seven seconds of an application.

Voice AI

Driver applies
AI voice call
Qualified applicant
Carrier CRM
Workflow overview

Vizco AI

Advisor & Founding Engineer · 2025–2026

Confidential files. Useful AI.

A legal-tech product that drafts and assembles case documents. I built the core agent system and took it from prototype to live client matters.

Read the engineering story
The problem
Confidential case files could not be sent directly to third-party language models.
What I built
Fail-closed redaction, an MCP tool platform over Microsoft 365, and versioned skills that let non-engineers extend agent behavior.
The outcome
A production workflow connected to firms’ existing document environments.

Document AI

Case documents
Redaction boundary
Agent workflow
Draft for review
Workflow overview

Vitaliti

Head of Engineering | Head of AI · 2023–2025

Context behind the recommendation.

A consumer healthcare startup where I built the engineering team and core architecture, and owned the AI roadmap.

Read the engineering story
The problem
Personalized guidance requires relevant context and a dependable product foundation.
What I built
A custom RAG system and knowledge graph, a contextual Health Assistant, and cloud infrastructure supporting the application.
The outcome
An engineering organization and AI product architecture built from the ground up.

Healthcare AI

User context
Knowledge graph
Retrieval
Health Assistant
Workflow overview

Where I've made impact.

From high-traffic advertising systems and fintech to leading teams building AI products.

2025 — PRESENT

CTO

call2action

AI-native recruiting platform for trucking: a driver-facing job marketplace wired to a carrier CRM, where an AI voice agent contacts a driver within seconds of applying.

Responsibilities & outcomes
  • Took an AI-native driver-recruiting platform from concept to a working two-sided product connecting carriers and drivers
  • Bet the product on AI-initiated contact: an AI voice agent reaches a driver within ~7 seconds of applying, replacing a recruiting model where callbacks take hours and most leads go cold
  • Designed the routing layer that delivers qualified, structured applicants into each carrier's pipeline without manual triage
  • Built the go-to-market data foundation: a pipeline over federal carrier records narrowing the national registry to 100,000+ targeted freight carriers
  • Made AI conversation quality measurable with an LLM classifier that scores every call, so recruiters act on outcomes instead of listening to recordings
TypeScriptNode.jsNestJSPostgreSQLGCPKubernetesAI Voice AgentsLLMOpenAI
2025 — 2026

Advisor | Founding Engineer

Vizco AI

Legal-tech company automating document-heavy legal workflows with an agent system that drafts and assembles client case documents end-to-end.

Responsibilities & outcomes
  • As founding engineer, built the company's core product and took it from prototype to production on live client matters
  • Unblocked commercial LLM use on confidential legal files by designing a fail-closed redaction architecture; without it, client documents could not be sent to third-party models at all
  • Made the product deployable inside firms' existing environments through an MCP tool platform over Microsoft 365, turning client onboarding from a data migration into a connection
  • Put capability delivery in non-engineers' hands with a versioned "skills" system, taking engineering off the critical path for every new agent behavior
  • Led reliability work that eliminated the failure modes taking the pipeline down under load
TypeScriptNestJSPythonAzureK8SAI AgentsMCPRAGPostgreSQL
2023 — 2025

Head of Engineering | Head of AI

Vitaliti

Led technology strategy. Built the engineering team and core architecture from scratch.

Responsibilities & outcomes
  • Leading technology strategy in alignment with business objectives
  • Managing AWS cloud infrastructure for scalability and reliability
  • Orchestrating containers with Kubernetes for operational efficiency
  • Implementing Infrastructure as Code with Terraform for consistent environments
  • Directing development in Node.js, TypeScript for back-end and React for front-end
  • Applying Domain-Driven Design (DDD) for aligning tech with business needs
  • Built the engineering team and development processes from scratch, including hiring, onboarding, and establishing engineering best practices
  • Designed and implemented the backend architecture of the core application with a focus on scalability, reliability, and performance
  • Optimized cloud infrastructure costs through architectural improvements and efficient resource management
  • Developed a custom RAG system and knowledge graph to power a personalized healthcare recommendation engine
  • Designed and launched an intelligent Health Assistant agent with contextual reasoning capabilities
AWSK8SPythonRAGTypeScriptReactTerraformPostgreSQLPrismaDDD
2021 — 2023

Technical Lead

Shopmonkey

System design, technology stack selection, and team leadership. Focused on performance optimization and efficient delivery.

Responsibilities & outcomes
  • Successfully designed and implemented the server-side architecture for a mobile application from scratch, increasing system performance by 40% and reducing maintenance costs by 25%
  • Built and led a high-performing team of 5+ engineers, ensuring timely delivery of features and improvements, leading to a 20% reduction in time-to-market
  • Optimized the technology stack selection process, reducing development time by 30% and improving system scalability, resulting in a 15% increase in user satisfaction
TypeScriptNestJSPrismaKafkaGraphQLMongoDBRedisElasticsearchAWSHelm
2019 — 2021

Senior Software Engineer

UAPAY

Developed a scalable fintech platform using microservice architecture to handle rapid business changes.

Responsibilities & outcomes
  • Designed and implemented over 60 microservices using domain-driven design, significantly enhancing system modularity and scalability, leading to a 25% increase in deployment efficiency
  • Played a key role in developing a highly scalable fintech startup, handling rapid business changes, which increased user adoption by 35% within the first year
  • Improved system performance and resilience by integrating advanced microservices architecture and AWS solutions, reducing system downtime by 40%
TypeScriptNode.jsPrismaDDDRabbitMQPostgreSQLMariaDBOracleRedisK8SJSON-RPC
2017 — 2019

Senior Software Engineer

Genesis Tech

Designed and developed backends for multiple applications, ensuring robust service-oriented architecture.

Responsibilities & outcomes
  • Designed and developed backends for 5 applications, ensuring robust and scalable service-oriented architecture, resulting in a 50% reduction in development costs
  • Created a comprehensive system for collecting and aggregating user behavior analytics, increasing product profitability by 27%
  • Successfully solved complex business problems through innovative software solutions, improving application performance by 30% and increasing user retention by 20%
TypeScriptNode.jsVueGraphQLWebSocketApolloPostgreSQLMySQLRedisRabbitMQAWS
2015 — 2017

Software Engineer

SmartyAds

Development and support of a highly loaded (10,000+ RPS) automated programmatic real-time trading system.

Responsibilities & outcomes
  • Developed both monolithic and microservice architecture for the platform
  • Created a custom database migration service for the platform
  • Initiated and introduced Continuous Integration technology into the project
Node.jsMySQLAerospikeClickHouseRedisMemcachedDockerOpenRTB
2012 — 2015

Android Developer & Frontend Freelance

Genii Studio & Freelance

Streaming media-player project with MVP architecture and freelance frontend development.

Responsibilities & outcomes
  • Used MVP architecture for streaming media-player project
  • Developed media player, Chromecast support, in-app purchases modules
  • Two years of frontend development experience (HTML, CSS, JS) as freelancer
JavaAndroid SDKGoogle CastHTMLCSSJavaScript

Built for the community.

From educational games to production RAG frameworks.

server-survival

6.4k GitHub stars Sep 2026

Tower defense game that teaches cloud architecture. Build infrastructure, survive traffic, learn scaling.

JavaScriptGameEducationCloud
View repository ↗

minima-aws

An open-source RAG framework connecting an S3 document pipeline, Qdrant retrieval and Amazon Bedrock.

PythonAWSRAGQdrantLLM
View repository ↗

stackcanvas

Live infrastructure canvas for AI coding agents over MCP — see and approve what an agent changes in your Terraform, locally.

TypeScriptMCPTerraformDevTools
View repository ↗

The Book.

At the intersection of machine learning and cognition.

How to Tell 5 from 6: How Machines Learned to Think

How to Tell 5 from 6: How Machines Learned to Think

A journey from the simple question of recognizing digits to the heart of modern AI. Discover how machines learn from examples—from pixels to perception—and understand the principles powering systems like GPT. No prerequisites, just a clear path to genuine understanding.

Find the book on Amazon

Ideas worth sharing.

Articles on architecture, AI, and lessons from the field.

ARTICLE

Why MikroORM is the Perfect Choice for Microservices

MEDIUM · SEP 2025
 ARTICLE

Building a Cloud-Based RAG System with Minima AWS

MEDIUM · APR 2025

Engineering notes.

Notes from building AI systems in production — strategy, architecture, and what actually matters.

BLOG

The Unreviewed Commit: What Code Review Was Actually For

JUN 23, 2026
 BLOG

Deterministic Governance in Autonomous Multi-Agent Systems

APR 7, 2026
 BLOG

Silicon Photonics and the Zero-Latency Inference Frontier

APR 7, 2026
Browse all articles ↗

Tools of the trade.

The tools behind my work in AI systems, application architecture and production infrastructure.

Languages

TypeScriptPythonJavaScriptNode.jsGoJavaSQL

AI & ML

LangChainRAGPyTorchTensorFlowOllamavLLMHugging FaceMCP

Frameworks

NestJSExpressApolloReactVue

Data & Messaging

PostgreSQLMongoDBClickHouseRedisKafkaRabbitMQElasticsearchQdrant

Cloud & DevOps

AWSAzureGCPKubernetesDockerTerraformHelmGitHub ActionsServerless

Architecture

MicroservicesDDDEvent-DrivenSOLIDRESTGraphQLWebSocket

Have a hard problem
worth solving?

Let’s talk about your product, the engineering challenges behind it, and where I can help.

Email me LinkedIn GitHub X Threads