Патрік

PATRIK SZAHAJDAK
SOFTWARE DEVELOPER

CONTACT ABOUT ME
[open contact info](look above in the "contact info" section) Full-Stack Developer with 2 years of
professional experience building and maintaining
[open contact info](look above in the "contact info" section) web applications and internal tools. Experienced
with Python, FastAPI, React, TypeScript,
Dunjaska Streda, TA, Slovakia PostgreSQL, SQL and REST APIs, with a strong
focus on backend development, databases and
github.com/RekteloPatrikOMGVERYBIG
problem solving.

EDUCATION CAREER HIGHLIGHTS
Software Developer SOFTEC SPOL. S.R.O, BRATISLAVA
SOFTWARE DEVELOPER | 2024-2026
It Step Academy, Kyiv Delivered 23 product features, including
15+ REST API endpoints and 18
2024 - 2026 React/TypeScript UI components,
working across Python, FastAPI,
PostgreSQL, React and TypeScript.
TECHNICAL SKILLS Improved backend performance by
reducing API response time from 1,676
Languages: Python, JavaScript, ms to ~500 ms and optimizing a
TypeScript, C#, C++ database query from 7 seconds to 2
seconds.
Improved page-loading performance by
Backend: FastAPI, Node.js, approximately 35% through frontend and
REST APIs backend optimizations.
Automated a previously manual workflow,
Frontend: React, Next.js, HTML5, reducing processing time from 35
minutes to 3 minutes and saving
CSS3, Tailwind CSS
approximately 7 hours per week.
Built and maintained database
Databases: PostgreSQL, SQL, functionality across 5 PostgreSQL tables
MongoDB, pgvector and implemented or optimized 8+ SQL
queries.
AI: Google Gemini API, Developed the Adaptive Complexity
feature largely independently,
embeddings, generative AI,
completing the implementation in
AI-assisted development approximately 3 days and reducing
related manual errors/actions by an
Testing: React Testing Library estimated 45–50%.
Collaborated within a 7-person squad as
Tools: Git, GitHub, Vercel, Render part of a broader 20–25 person
engineering organization, typically
completing 2–3 tasks per sprint.

PROJECTS
N5DEAL — M&A MARKETPLACE PROTOTYPE NEXUS — AI-POWERED PLATFORM

Built and deployed a full-stack M&A and Developing a full-stack AI-powered
financial-assets marketplace with 3 role- developer platform using Next.js, React,
based experiences — Buyer, Seller, and TypeScript, Supabase,
Platform Manager — across 19 page PostgreSQL/pgvector and Google Gemini
routes using Next.js, TypeScript, API.
Supabase and PostgreSQL. Built 17 React/Next.js components across
Implemented 12 API endpoints and the first 2 application pages and
approximately 12–15 end-to-end user implemented 3 API routes during the first
flows, including asset creation and month of development.
publishing, buyer discovery, seller Integrated AI functionality with typical
contact, messaging, profile management response times of approximately 1–2
and participant moderation. seconds and tested the system with
Designed a PostgreSQL data model with 100+ AI requests during development.
5 relational tables and 7 foreign-key
KNOWLEDGEAI — AI ASSISTANT
relationships covering user profiles,
buyer interests, marketplace assets, Built a document-grounded AI assistant
contacts and messages. supporting 4 source types — PDF, TXT,
Developed a rule-based natural-language DOCX and URLs — for creating
smart search supporting 6 asset search searchable knowledge bases.
parameters, 5 filtering criteria, 4 sorting Tested document ingestion with files up
modes, 6 industries and 5 countries, to 124 pages / 70,000+ words, using
without relying on an external LLM API. retrieval chunks of approximately 500
Built 10 user-facing forms, including a 14- tokens each.
field buyer acquisition profile, and Achieved document indexing times of
implemented 10+ write operations across approximately 2–7 seconds and AI
marketplace, messaging and moderation response times of 1–3 seconds in
workflows. development testing.
Created responsive layouts for at least Implemented retrieval-based answering
16 core pages and shipped the working with source grounding, achieving
prototype to Vercel with persistent approximately 8/10 successful source
Supabase-backed marketplace data. retrievals in initial tests.