Jakub Towarek
I'm an iOS Engineer with 5 years of experience crafting elegant solutions for Apple platforms. I blend technical knowledge with user empathy to transform complex requirements into intuitive interfaces that millions of users interact with daily.
Book a meeting | Mail | Mobile | GitHub | LinkedIn
π‘ Technologies
- π
SwiftUIUIKit
- πͺ’
ConcurrencyCombineRxSwift
- πΊοΈ
GitGit FlowSemantic VersioningChangelog
- π§¦
Swift Package ManagerCocoaPodsCarthageMint
- π
RESTWebSocketsgRPCGraphQLOpenAPI
- π¦Ί
GitHub Actions w/ Self-Hosted RunnersBitrise
- π€
Firebase Remote ConfigCrashlyticsGoogle AnalyticsAmplitude
- π§³
App Store ConnectTestFlightApple Business ManagerJamf
π₯· Experience
Wisent | Founding Engineer
May 2025 ~ Present
- built Wisent's native iOS app from zero to one, including on-device LLM, image-generation, and TTS inference with MLX
- provided live user support via Discord, closing the loop between user feedback and product iteration
- build the inference-time runtime that exposes concept directions for read, write, and clamp operations
- build and maintain the contrastive-pair concept-dataset pipelines and per-layer activation capture spanning HuggingFace transformer stacks at sizes from 1B to 405B
- architect Wisent's multi-cloud infrastructure: European HPC allocation for training, a Python/FastAPI backend on AWS (Lambda, EventBridge, EC2, S3, CloudFront), and the app.wisent.ai web application
- re-architected backend from a monolithic A100-coupled service into independently deployable CPU API, inference, and image services, removing model-load from routine releases and reducing deploys from 36-44 min to ~8 min (~78-82% faster) while eliminating routine deploy-time A100 usage (~36-44 A100-min saved per deploy; 100% GPU deploy-cost reduction)
- ported on-device speech/LLM inference stack to iOS by implementing NeuCodec in Swift/MLX and extending Swift Transformers/CoreML tokenization support, covering a 24kHz / 50Hz neural audio codec, 2048->1024 projection, 12-layer / 16-head Transformer, ISTFT decoding, safetensors weight loading, and Python-reference parity tests, enabling server-free real-time TTS/voice inference on Apple devices
- led cross-platform, multi-layer safety/feed moderation across iOS, Android, and web by migrating home/search/category feeds to server-filtered RPCs, adding automated NSFW/pending gating plus an admin moderation panel with calibration tooling, and replacing duplicated client filters with consistent policy enforcement across 3 product surfaces
SumUp | iOS Engineer, Online Store Tribe
February 2023 ~ February 2025
- own the full development cycle for key merchant-facing features, including order management used by over 100,000 merchants
- fix and contribute to the design system, including parts written in Obj-C and consumed across 5 tribes
- take part in design reviews, providing actionable feedback and suggesting "quick wins" for UI / UX designs on iPhone and iPad
- start an app-wide migration to SwiftUI
- monitor app stability via Crashlytics, feature performance via Google Analytics, and overall system health via traces on Honeycomb; crash-free sessions improved from 98.7% to 99.6%
- effectively collaborate within Spotify's Agile model (tribes, squads, chapters) to deliver features with global impact, like Customer Directory and integrations from different tribes
- research spikes and summarize solutions, for example implementing a rich text editor with quill.js via WebView as a less resource-intensive alternative to creating a native solution
- improve OpenAPI definitions to correctly express behaviors and types, making iOS client generators less error-prone
Dispo | iOS Engineer
January 2022 ~ February 2023
- independently implement new features in an existing codebase
- redesign, optimise and expand capabilities of real-time image processing system
- work closely with design team on the release of image editing suite and introduction of new camera effects
- maintain stability by monitoring and prioritising crashes and non-fatal errors
- improve performance and readability of asset caching mechanism using async await pattern
- prepare features to be released as part of A/B experiments in order to minimise potential negative outcomes
- create and upload new app versions to the App Store Connect
- progressively refactor The Composable Architecture to MVVM with SwiftUI, Combine and async await
- experiment with disparity-based depth measuring techniques for advanced 3D photo effects
Netguru | iOS Research and Development Lead
August 2021 ~ January 2022
- coordinate AR virtual try-on app development
- filter iOS trends and news to select
- advise and provide feedback on blog posts and presentations
- align team members' needs with research topics
Netguru | iOS Developer
January 2020 ~ January 2022
Design System For Food Delivery Company
- design common interfaces for Android, iOS, and Flutter
- write extensive documentation
- cover components with screenshot tests and UI Tests
Voice Assistant App For Easy CRM Access
- lead architecture design with
SwiftUIandCombine - implement on-device speech transcription
- interact with
serverlessbackend viawebsockets - advance communicated project value to stakeholders through analytics
- monitor and prioritise crashes and errors
IoT App For Startup
- add new features in obscure codebase
- fix critical security vulnerabilities in
Firebase Authentication
Netguru | Junior iOS Developer
October 2019 ~ January 2020
π‘ Open Source
wisent
Co-author and maintainer of the reference representation-engineering library for text LLMs: 342 GitHub stars, 32 forks, over 200,000 downloads, latest release v0.11.43 (2026-05-18), continuous releases since 2024.
SwiftOLED & SwiftGFX
Two Swift libraries developed while tinkering with Raspberry Pi and SwiftOnArm. I wanted to use Swift, but there wasn't a library that would let me display data on an external OLED display, so I created two - one for drawing a pixel buffer on the screen, and a second graphics library capable of rendering text to a pixel buffer. Both have nice READMEs. I even got some contributions.
π’ Languages
I am fluent in English - I hold an IELTS C1 certificate.
π Education
After finishing technical high school and passing my IT Specialist certification, I started an Applied Computer Science major at Newcastle University in the UK. Shortly after, I got accepted for a junior iOS developer position, so I dropped out.