• Файл

Sergey

Web software engineer

Місто проживання:
Кривий Ріг
Готовий працювати:
Дистанційно

Контактна інформація

Шукач вказав: ТелефонЕл. поштуМесенджер

Прізвище, контакти та світлина доступні тільки для зареєстрованих роботодавців. Щоб отримати доступ до особистих даних кандидатів, увійдіть як роботодавець або зареєструйтеся.

Завантажений файл

Версія для швидкого перегляду

Це резюме розміщено у вигляді файлу. Ця версія для швидкого перегляду може бути гіршою за оригінал резюме.

SERHII FILIN
Senior Software Engineer | Full-Stack, AI
[відкрити контакти](див. вище в блоці «контактна інформація») | [відкрити контакти](див. вище в блоці «контактна інформація») | [відкрити контакти](див. вище в блоці «контактна інформація») | t.me/f_1_1_1_n

SUMMARY
TypeScript developer with 7+ years of experience. Builds products end to end, from database schema to interface, with real-time
data at scale, rendering performance and incremental framework migrations as the core strengths. Also ships AI features to
production, with failure handling, measured quality and predictable cost.

SKILLS
Core: TypeScript, JavaScript, Python, HTML, SASS, Tailwind CSS
Front-end: Angular, Signals, standalone components, new control flow, OnPush change detection, RxJS, NgRx, NGXS, Angular CDK,
Angular Material, PrimeNG, SSR and hydration, Nx monorepo, Storybook, React, TanStack Query, Redux Toolkit, Zustand, React Hook
Form, Vue.js, Next.js (SSR/ISR)
Back-end: Node.js, NestJS, Express.js, REST, GraphQL, WebSocket, PostgreSQL, MongoDB, Redis, TypeORM, Prisma, BullMQ,
RabbitMQ, migrations and indexing, JWT, OAuth2, RBAC, OpenAPI/Swagger, Stripe webhooks
Data and maps: Chart.js, ECharts, canvas and WebGL rendering, Google Maps, MapLibre, Leaflet, GeoJSON, marker clustering
AI and LLM: Anthropic Claude API, OpenAI API, RAG, embeddings, vector databases, agentic workflows, function/tool calling,
structured output validation and guardrails, streaming (SSE), prompt caching, evaluation harnesses with golden sets, LLM-as-a-judge
gating, multi-vendor routing and failover, token cost instrumentation
Performance: Core Web Vitals, code splitting, lazy loading, list virtualization, bundle optimization
Tooling: Docker, CI/CD, Webpack, Sentry, Jest, Cypress, Playwright, Git

PROFESSIONAL EXPERIENCE

TrackEnsure | Software Engineer | 09/2025 - Present
Fleet telematics and ELD compliance platform for US trucking
Angular 22, TypeScript, RxJS, WebSocket, MapLibre, Python, pytest, LLM evaluation harness
• Designed one abstraction layer over three map SDKs (Google Maps, MapLibre, Leaflet), keeping vendor-specific code confined to
the adapters.
• Scaled the fleet map to 10,000+ tractor and trailer markers at 60 FPS with clustering, live telemetry, playback of historical tracks
and a planned-vs-actual view that surfaces route deviations, unscheduled stops and late loading.
• Built the HOS timeline against the 11-hour driving limit, the 14-hour window and the rolling 70/8 cycle, with violations marked
on the timeline.
• Replaced manual quarterly IFTA reconciliation with a one-click export covering mileage and fuel across 58 jurisdictions, flagging
discrepancies before filing.
• Shipped the fleet metrics dashboard with selection linked between charts and the map, so a spike in idle time or fuel use
resolves to specific trucks and route segments.
• Built a provider-agnostic LLM layer for extracting structured fields from driver-submitted fuel receipts and shipping documents,
with per-vendor adapters, failover and one telemetry format across vendors (latency, error rate and cost per request), running in
production.
• Built an internal LLM-as-a-judge release gate on my own initiative, adopted by the team as a CI check: every build scores against
a held-out golden set and must clear a fixed pass threshold, failing quality regressions the day they land.
• Analyzed session cost data and turned the result into a production routing rule: subtasks moved to a cheaper model tier now
produce ~34% of output for ~15% of spend, a 3x difference in blended cost per token.
Cryptology Key | Full-Stack Engineer (TypeScript, AI) | 03/2024 - 08/2025
Education platform for traders: courses, community and market analytics
React, TypeScript, NestJS, PostgreSQL, MongoDB, GraphQL, Python, SSE, RAG, vector store, prompt caching
• Ran courses, community and market analytics end to end, owning both the front end and the back end.
• Consolidated courses, progress and analytics behind a single GraphQL endpoint, removing multiple round trips and over-
fetching.
• Integrated Stripe for subscription-gated content, with webhook handling that never charges or grants access twice on a
redelivered event.

• Built scheduled market data ingestion in Python into PostgreSQL, replacing heavy queries at request time with composite
indexes and precomputed slices.
• Streamed assistant responses over SSE with incremental rendering, mid-generation cancellation and history preserved across
reloads. In-flight answers recover after a dropped connection.
• Built the RAG layer with permissions enforced inside the retrieval query itself, so results are scoped at the source. Deleting a
document also removes its embeddings.
• Declared every tool for function calling by schema and validated structured model output before it reached business logic, with
an approval gate on any action that moves money or changes permissions.
• Moved long-running agentic jobs onto a queue with state in the database, idempotent steps and failure-aware retries, so a
worker crash resumes from the last completed step.
Loyalty System (LSbonus) | Front-End Developer | 02/2023 - 02/2024
B2B SaaS loyalty platform for retail chains
Angular, TypeScript, RxJS, canvas rendering
• Built the merchant CRM covering bonuses, multi-balances, coupons, gift cards and campaigns, all configurable per tenant
without a code change or release.
• Virtualized the customer base view for 100k+ rows with server-side filtering and segmentation. Segment state lives in the URL,
so a segment can be shared as a link.
• Moved chart rendering from SVG to canvas, which kept pan and zoom responsive on retail datasets where the SVG version
stalled.
• Cut the main bundle by 40% by dropping moment.js and switching lodash to per-method imports.
IDAP | Front-End Developer | 05/2019 - 12/2022
Fintech, medtech and retail products
Angular, TypeScript, RxJS, NgRx, Chart.js, Plaid.js
• Designed the real-time data layer over WebSocket with reconnect, buffering and deduplication, eliminating data loss on network
switches.
• Cut LCP from 4.1s to 1.8s through critical CSS, font preloading and lazy hydration of below-the-fold blocks.
• Replaced per-message redraws with requestAnimationFrame batching for high-frequency streams, holding frame rate steady
under sustained bursts.
• Led the incremental Angular 8 to 14 migration, six majors including Ivy, with no pause in feature delivery.
• Set the front-end engineering baseline (standards, Git workflow, CI/CD) and ran the internal training behind it, cutting ramp-up
for developers rotating between projects.
• Reviewed work across teams and settled API contracts with back-end before implementation.

EDUCATION
Kyiv National Economic University | BSc, Computer Engineering | 2019

Схожі кандидати

Усі схожі кандидати

Кандидати у категорії


Порівняйте свої вимоги та зарплату з вакансіями інших підприємств: