Available for work

Senior Frontend Engineer / AI Systems Builder / Kazakhstan

Islambek Chynybekov

I build high-discipline interfaces and AI-augmented product systems that feel precise, fast, and unmistakably production ready.

View Work Get in Touch
  • 5+ years shipping digital products
  • 30+ projects delivered across fintech and AI
  • AI × UX frontend systems with practical automation
About Me

Building the bridge between
AI & great UX

I am a Senior Frontend Engineer based in Kazakhstan with 5+ years in FinTech and product delivery. My strongest work happens where interface quality, system thinking, and execution discipline have to coexist under real business pressure.

At Halyk Finservice I lead a frontend stream and introduced AI agents into review, documentation, and testing workflows. The point is not novelty. The point is raising team throughput without lowering engineering standards.

I focus on building calmer systems: clearer interfaces, cleaner delivery loops, better architecture decisions, and AI that actually removes friction instead of adding noise.

Frontend Engineering AI / LLM Integration RAG Architecture FinTech Team Leadership Vue / React Swift / macOS
Operator Snapshot
Current role Frontend Lead / AI Systems Builder
Primary domain FinTech, internal platforms, delivery systems
Core edge Turning complexity into cleaner execution
Working style High standards, low noise, production-first decisions
How I work
  • System thinking before feature clutter.
  • Performance as a product trait, not a polish pass.
  • AI used for leverage, never for noise.
Skills & Tools

Technologies I work with

Frontend

  • Vue 3 / Nuxt 3
  • React / Next.js
  • TypeScript
  • Tailwind CSS / SASS
  • Capacitor / Ionic
  • Tauri Framework

AI & ML

  • LLM Integration (GPT, Claude)
  • RAG Architectures
  • AI Agents & Automation
  • Vector Embeddings (pgvector)
  • Apple NLEmbedding / Vision
  • Hybrid Retrieval (BM25 + Vector)

Backend & Data

  • Python / FastAPI
  • Swift 5.10 / macOS
  • Node.js / Nest.js
  • PL/SQL / PostgreSQL
  • REST API / GraphQL
  • Blockchain (Web3)

Tools & DevOps

  • Git / GitHub / CI-CD
  • Docker / Railway
  • Webpack / Vite
  • AppStore & Play Store
  • Figma / UI Design
  • Unit / E2E Testing
Experience

My professional journey

Jan 2024 — Present

Team Lead Frontend

Halyk Finservice · Almaty, Kazakhstan

Leading a stream of developers in Kazakhstan's largest bank subsidiary. Pioneered AI agent integration into the dev pipeline — LLMs now perform automated code reviews, generate documentation, and write unit & E2E tests, cutting manual overhead significantly. Responsible for architectural reviews, task decomposition, and production deployments.

Vue 3TypeScriptLLM AgentsCI/CDTeam Lead
Jul 2023 — Dec 2023

Frontend Developer

Halyk Finservice · Almaty, Kazakhstan

Refactored legacy codebases and built a unified UI component library that standardized development across all projects. Mentored junior developers and optimized frontend-backend communication protocols.

Vue 3UI LibraryMentoring
Aug 2022 — Jul 2023

Software Developer

Kaspi.kz

Optimized high-load internal CRM systems for Kazakhstan's leading super-app. Collaborated with data teams to improve CRM dashboard visualizations. Migrated legacy PL/SQL logic to modern web architectures, enhancing performance at scale.

Vue 3TypeScriptPL/SQLData Viz
Feb 2022 — Jun 2022

Full-stack Developer

Air Astana · Almaty, Kazakhstan

Designed and built a system for optimizing aircraft de-icing processes. Analyzed operational workflows, designed the project architecture, and managed team members through delivery.

Vue.jsQuasarFastAPIPython
Aug 2020 — Aug 2022

Frontend Developer

Adamar Solutions · Almaty, Kazakhstan

Built and shipped multiple cross-platform apps to the AppStore and Play Store. Wrote unit, integration, and e2e tests. Worked across a broad stack of modern frontend and hybrid mobile technologies.

Vue 3Nuxt 3CapacitorIonicTauriTailwind
Jan 2019 — Feb 2020

Support & Video Creator

Morrison · Almaty, Kazakhstan

Recorded 60+ tutorial videos on JavaScript, Web Development, and React.js. Delivered 50+ lessons with structured presentations and supervised groups of up to 20 students.

JavaScriptReactTeaching

Want the full story?

Download my complete CV for a detailed overview of my experience and skills.

Download CV
Projects

Selected case studies

Case Study 01

Overlayer AI — Ambient context engine for macOS developers

A native macOS app that continuously captures screen context via Vision OCR, accessibility trees, and clipboard — surfacing contextual hints in a frosted-glass overlay during natural developer pauses. Zero telemetry. Local-first.

  • Hybrid search: Apple NLEmbedding (vector) + FTS5 (keyword) + reciprocal rank fusion — all on-device.
  • Complexity-routed LLM dispatch: simple → Haiku, complex → Sonnet, privacy-sensitive → local Apple Foundation Models.
  • MCP server exposes context to Claude Code and external agents. Swift 5.10 + GRDB SQLite. Apache 2.0.
GitHub
Platform Native macOS 14+

Swift + Vision framework + NLEmbedding. No Electron, no web shell.

Architecture Privacy-first, local compute

All embeddings and search run on-device. LLM routing by privacy sensitivity.

Swift 5.10Vision OCRNLEmbeddingGRDB SQLiteMCP
Case Study 02

HalykMarket AI delivery system — Enterprise frontend at bank scale

Built a practical AI layer inside the engineering workflow at Halyk Finservice — Kazakhstan's largest bank subsidiary. Automated review support, documentation generation, test drafting, and cleaner delivery rituals.

  • Inserted AI agents where manual repetition was highest and quality drift was most expensive.
  • Framed the system as delivery leverage, not demo-ware. Architectural review and final judgment stayed human-led.
  • Ghost Mode: zero AI attribution in external artifacts (MRs, commits, emails, Jira) — enforced systematically.
Context
Focus Velocity without entropy

Reduce overhead while raising consistency across the team.

Core stack Vue 3, TypeScript, LLM agents

Operational AI applied directly to delivery flow.

Review automationDocumentationTestingGhost Mode
Case Study 03

tg-moder — Production RAG moderation system at $2/mo

Production Telegram moderation bot for a 17K-member community. 3-pass hybrid retrieval (pgvector HNSW + tsvector BM25 + ILIKE) with RRF k=60. Deployed on Railway with shadow-before-live discipline.

  • 10-stage cost cascade: cache → FAQ → route → RAG → LLM → judge. $2/mo operational cost with ~67% FAQ hit rate.
  • Confidence gate 0.65, relevance floor 0.15, monthly budget hard cap ($5) checked BEFORE embed call.
  • Shadow mode first: logged every response before enabling live replies. Replay fixtures + pre-deploy checks.
GitHub
Retrieval 3-pass hybrid + RRF k=60

pgvector HNSW · tsvector BM25 · ILIKE. OpenAI text-embedding-3-large (3072d).

Cost $2/mo operational

Budget hard cap enforced pre-embed. LLM cost cascade, not brute force.

pgvectorBM25RRF k=60RailwayPython
Case Study 04

Shipping under pressure — hackathon wins and App Store delivery

Two hackathon 1st places and sustained cross-platform delivery to App Store and Play Store. Speed of execution without architectural shortcuts.

  • Savabit Mining hackathon — 1st place: blockchain/mining operations MVP, full product cycle under 48h.
  • Ioka.kz × Mastercard hackathon — 1st place: payment platform MVP with live transaction flows.
  • Adamar Solutions: shipped multiple apps to App Store + Play Store using Vue 3 + Nuxt 3 + Capacitor + Ionic + Tauri.
LinkedIn
Wins 2× hackathon 1st place

Named events, verifiable on LinkedIn. Product speed without cutting engineering corners.

Platforms Web · iOS · Android · Desktop

Sustained cross-platform delivery — not prototypes, shipped to stores.

FinTechPaymentsBlockchainCapacitorTauri
Education

Learning never stops

2019 — 2023

Bachelor's in IT

Almaty Management University (AMU)

Degree in Information Technologies. Built a foundation in software engineering, algorithms, and data structures. Actively participated in hackathons and student tech competitions.

Cisco · 2022–2023

Networking & Security

Cisco Networking Academy

Earned Cybersecurity Essentials (x2) and CCNA: Introduction to Networks certifications — strengthening understanding of secure, scalable system architecture.

Ongoing

AI Engineering

Self-directed Research

Actively researching LLM agent frameworks, RAG architectures, and cutting-edge model implementations. Building hands-on AI projects and studying emerging tools in the AI engineering space.

Evidence

What this work is really about

Delivery

I optimize for throughput that still feels deliberate.

Speed matters only when architecture, review quality, and product clarity survive it.

AI

I use AI as leverage inside the workflow, not as decoration on top of it.

Automation belongs where teams repeat expensive manual work and where consistency matters.

Leadership

I make systems calmer for the people shipping them.

That means cleaner decisions, clearer interfaces, better decomposition, and less operational chaos.

Fit

My best environment is ambitious product work with real complexity and real stakes.

FinTech, internal platforms, AI-enabled tooling, and teams that care about execution quality.

Contact

Let's build something
amazing together

theluruous@​icloud.com
Email → LinkedIn → GitHub → Versions →

© 2026 Islambek Chynybekov