
Ferhat Abdullahoglu
Verified Expert in Engineering
iOS Developer
Istanbul, Turkey
Toptal member since April 12, 2022
Ferhat is a senior full-stack engineer with 13 years years of experience who builds AI-powered products end-to-end: native iOS apps in Swift, SwiftUI, and TCA; Python/FastAPI and Supabase backends; and the LLM layer in between (agents, RAG, vision, and generative media). He has shipped several AI apps to the App Store, built a RAG-based ESG risk monitoring tool used by major Turkish enterprises, and founded a meditation app that was later acquired.
Portfolio
Experience
- Swift - 9 years
- iOS - 9 years
- SwiftUI - 6 years
- The Composable Architecture (TCA) - 4 years
- OpenAI - 3 years
- Large Language Models (LLMs) - 3 years
- Python - 3 years
- FastAPI - 2 years
Work Experience
Lead Engineer, AI Products
GPT Genius Lab
- Built a reusable product platform: a modular SwiftUI/TCA iOS template and a FastAPI/Supabase back end on Fly.io, with scripts that provision Apple, Firebase, Sentry, RevenueCat, and AppsFlyer for each new app.
- Shipped Glipath, a local-first GLP-1 tracker using SQLite, HealthKit, widgets, and Live Activities, with AI food logging from photos, barcodes, and voice.
- Released Damson, Namora, AI TLDRR, and Meditate Mate AI on the App Store, covering LLM generation with Agno and GPT-4o, CloudKit sync, share extensions, and Whisper transcription.
- Built baby-face prediction with FLUX on Modal serverless GPUs, plus ComfyUI workflows on RunPod for app features and ad creatives.
- Built chess-position recognition from photos with a vision LLM and an RF-DETR piece detector side by side, a referee model for disagreements, and a benchmark harness every change must beat.
- Built agentic research systems with Playwright readers and an LLM judge whose quotes are verified word for word in code before entering an evidence graph.
- Created Cortex, an MCP memory server for AI coding agents with Postgres/pgvector hybrid search and multi-tenant row-level security.
- Implemented growth tooling: RevenueCat paywalls, Mixpanel, PostHog, and AppsFlyer analytics, App Store optimization, and Meta and Google Ads API automation.
iOS Developer
Agrofocal Technologies Inc.
- Implemented the background upload logic and integrated it with the existing logic.
- Built the background upload solution into the existing solution without requiring any other changes in the codebase, as Alamofire was being used to handle API calls.
- Implemented recovery logic so that if an upload gets interrupted for any reason, it would be restarted and continued from where it was left off upon the subsequent launch of the app.
Co-founder & CTO
Captanomy
- Built an AI-based, RAG-assisted ESG risk monitoring tool: crawlers gather public pages and documents, Llama 3 on Bedrock filters ESG content, and LangChain/OpenAI pipelines answer risk questions. Used with some of Turkey's largest companies.
- Achieved three working models, managed with the same codebase from a single source: a white-label solution with a custom domain the client chooses and a partnership model where the client manages their partners' journey from a custom subdomain.
- Partnered with Turkey's greatest hub for SMEs as the single tool for carbon emissions calculation and reporting.
- Oversaw the technical stack transformation from a bare-bones POC to an industry-grade carbon accounting platform built on Laravel, React, and Docker.
Senior Swift Developer
Superbright Studios LTD
- Developed a project to support multiple operating system targets, mainly for macOS, iOS, and visionOS.
- Integrated with the existing team of graphic designers and developers who were already experienced with platforms like Oculus to bring their product onto Apple platforms.
- Oversaw the successful release of the public beta, which was also reviewed by a team from Apple in person.
- Delivered a high-quality product with a scalable and testable codebase much faster than the planned timeline, which made the engagement a success overall.
- Integrated well with the existing team to deliver the product on time and meet all public beta requirements.
iOS Developer
RD338 Design, LLC
- Spun up an MVP for a standalone watchOS app within a minimal time.
- Integrated App Intents for using the app without the user having to launch the app.
- Fed user tasks to App Intents so that the next time the same task needs to be executed, it wasn't required to launch the app, and it was enough to talk to Siri.
Consulting Senior Swift Developer | Apple Watch Tech/HR
I Love Your Stories, LLC
- Leveraged Apple background modes to have the watch listen and react to the updates from the iPhone app even when the wrist is down or the screen is off.
- Created a base project architecture using SwiftUI and TCA for the client developer to continue the development after our engagement ended.
- Used codebase SPM while rearchitecting to allow shared logic between watchOS and iOS targets.
iOS Developer (Swift/Firebase)
Castro Labs, Inc
- Developed two separate apps with the client, one in the dating industry and another in the social networking industry.
- Worked on a dating app that offered similar features to the most popular dating apps available today. From its early stages as an MVP, I contributed to its development until it evolved through several versions on the App Store.
- Built using the composable architecture (TCA) and SwiftUI. The app followed micro apps architecture, encapsulating individual features within Swift packages.
- Improved compilation speed and enhanced the reliability of SwiftUI previews using a modular approach. The architecture resulted in a highly flexible, testable, and efficient codebase.
- Worked on profile discovery, where users were presented with profiles of other individuals based on specific parameters, facilitating potential matches based on their preferences.
- Contributed to matching, where for communication to occur, both individuals were required to match each other mutually, indicating mutual interest.
- Worked on liking and commenting, where users could express their interest in other individuals by liking their photos and leaving comments, fostering engagement and interaction.
- Worked on messaging, where a messaging feature was implemented, enabling matched individuals to communicate privately within the app (implemented using GetStream).
- Contributed to a social media app that revolved around communities managed by community hosts. The users could seek and join these communities, which facilitated both group chats and direct messaging among community members.
- Used Swift structured concurrency throughout the codebase. Here is a replica of the API client I built for this project: github.com/ferhatabd/swift-api-client.
Co-founder | iOS and PHP Developer
Crea Meditation
- Developed the second Turkish mindfulness-based meditation app used by hundreds of thousands of users.
- Got the app featured on the App Store multiple times as it is a unique project in the market, allowing users to create their meditations following their preferences.
- Sold the app to a publishing company, which was translated into eight languages.
- Worked on HLS Streaming of the content on the watch and the iPhone.
- Worked on the back-end application with PHP and Laravel for content management and user database.
- Used AWS MediaConvert and HLS Stream, the back end would upload all the content to an S3 bucket. MediaConvert would then pick up the newly uploaded content and convert them to an M3U8 playlist for HLS Streaming.
- Complemented the app by a companion watchOS app through which the users can access and engage with their favorite mindfulness sessions and any additional content they have purchased.
- Worked on the personal affirmation meditation model, leveraging AWS Lambda and MediaConvert; pre-recorded sound recording files were intelligently stitched together, culminating in creating a unique and personalized affirmation meditation for the user.
- Utilized Web2App, a React landing page that allowed users to answer a series of questions based on which a personalized meditation list was created for the user.
Co-founder | iOS and PHP Developer
Res Pilates App
- Actively contributed to the development of the Res Pilates App. As part of this endeavor, I focused on building iOS and watchOS applications. https://apps.apple.com/app/id1495548426.
- Worked the app that served as a reliable and user-friendly companion for pilates workouts, with the added benefit of seamlessly integrating with the Apple Watch to track active sessions and workout history.
- Built the app using PHP and Laravel. In addition to authentication, the server application manages user perks based on referrals, creating a system that incentivizes engagement and growth.
- Worked on the server application that handles content management. It interfaces with Contentful, a content management system (CMS), to facilitate the app's organization and delivery of content.
- Worked with multiple ways to interact with the workout video player, interruptions can occur from various sources, including the iPhone, Apple Watch, TV remotes, and the operating systems of each respective device.
- Worked with Siri Voice Control; users can control video playback and initiate any workout by commanding Siri on either the watch or iPhone. This integration with Siri leverages the capabilities of the Intents and SiriKit frameworks.
Application Manager
B&R Industrial Automation
- Created and managed company-wide reusable software packages, increasing the team's efficiency and the quality of the products delivered through standardized and thoroughly tested software components.
- Built an efficient and successful team of talented engineers.
- Managed the team resources and monitored the development process to deliver successful products that fulfill clients' requirements and needs.
Application Engineer
B&R Industrial Automation
- Used C++ to develop a solution for a specific machine that enables users to program the machine movement with a predefined set of commands and rules on the human-machine interface (HMI) with a text editor.
- Developed the first-of-its-kind software for a 25-axis glass-processing CNC center. It was delivered and commissioned in many countries.
- Created software for an in-house standardized software package for a specific machine type used in many diverse company branches on different continents.
Project History
Glipath: GLP-1 & Peptide Tracker
https://apps.apple.com/app/id6788376667• Designed a local-first iOS 26 app with TCA and SQLite in an App Group database, so health data stays on the device and is shared with widgets and a Live Activity.
• Integrated HealthKit and built AI food logging from a photo, a barcode, or a voice note.
• Built a stateless FastAPI and Supabase back end for food AI and insights, with a medication catalog that cites pharmacokinetic sources.
• Wired in RevenueCat subscriptions, PostHog, Mixpanel, and AppsFlyer analytics, and ran the paid acquisition setup.
• Backed the codebase with extensive Swift and pytest test suites.
Damson: PCOS Meal & Symptom Tracker
https://apps.apple.com/app/id6796829970• Built a local-first SwiftUI and TCA app with SQLite storage and CloudKit sync for premium users.
• Implemented meal capture by photo, barcode, or manual entry, plus symptom tracking.
• Built a server-side reflection engine in Python/FastAPI that uses LLMs and embeddings to turn a user's logs into personal, evidence-aware insights.
• Covered the iOS app and the back end with large automated test suites to keep a fast release cadence safe.
Kibitz: Chess Social App
https://kibitzer.gg• Built chess-position recognition from photos that runs a vision LLM and a dedicated piece detector (RF-DETR) side by side, with a referee model settling the squares they disagree on and uncertain squares flagged instead of guessed. Every change has to beat a benchmark harness of real positions.
• Designed the back end in FastAPI and documented the architecture with ADRs, a threat model, a load-test report, and a launch-readiness checklist.
• Built the native SwiftUI app and an Astro web app on Cloudflare Workers, plus a Lichess bot integration.
• Backed the system with thousands of automated tests across Python and Swift.
Namora: AI Baby Names & Numerology
https://apps.apple.com/app/id6761501309• Built LLM-driven name generation with Agno and GPT-4o, personalized to the parents' preferences and numerology.
• Built baby-face prediction with FLUX models on Modal serverless GPUs.
• Developed the SwiftUI and TCA app on a FastAPI and Supabase back end, with RevenueCat subscriptions.
• Ran App Store optimization and paid acquisition experiments after launch.
AI TLDRR: Save for Later With AI
https://apps.apple.com/app/id6759262513• Built share and action extensions so users can save content from anywhere on iOS.
• Built a Python/FastAPI pipeline that extracts content, transcribes audio with Whisper, and summarizes it with LLMs.
• Implemented semantic search and retrieval using OpenAI embeddings, Qdrant, and a Memgraph knowledge graph.
• Shipped to the App Store with subscriptions and analytics.
Cortex: Memory Layer for AI Coding Agents
• Exposed brief, search, remember, and check tools over both MCP and REST.
• Implemented hybrid retrieval in PostgreSQL that combines full-text search and pgvector embeddings with reciprocal rank fusion.
• Designed multi-tenant isolation with Postgres row-level security.
• Added verification checks so agents confirm facts against the codebase before acting on stored memories.
Bent | Queer Communities
OOMF! | Gay and Queer Messaging
Meditate Mate AI
https://apps.apple.com/app/id6737718272The server app used GPT-4o and Llama 3.3 to handle conversations with the user and generate scripts. The iOS app was developed with SwiftUI. The app is one of a kind, and I delivered the project successfully.
Mindfulness Meditation App
https://apps.apple.com/app/id1509910196This was made possible by AWS and our custom back-end systems working together. The back-end processes the meditation request, which sends a request to AWS Elemental MediaConvert, using the AWS API Gateway to create a new meditation session. Finally, the app gets this information from the back end to serve the content to the user. We also implemented the Web2App concept, a feature that administers a quiz on https://premium.crea.care and allows the creation of a specific set of meditations based on the answers.
RAG Chatbot for Public ESG Data
• Crawling a company's public webpage
• Using Llama over Amazon Bedrock to identify which links are related to ESG
• Scraping text content, downloading PDF docs, and storing everything on AWS with a combination of DynamoDB and Amazon S3
• Generating embeddings using text-embedding-ada-002 from OpenAI
• Storing resulting vectors in Pinecone
• Switching the context based on the selected company and answering questions through RAG using GPT-4
Swift API Client
https://github.com/ferhatabd/swift-api-clientIt's especially helpful as it still uses Alamofire, which is a very well-known framework, in its backbone and enables defining fully typed routes and executing the client from within an async scope
Pilates Workout App
https://apps.apple.com/gr/app/res-pilates/id1495548426FAStoryKit Open Source Library
https://github.com/ferhatabd/FAStoryKitIt makes use of Generics to support different content types.
MindfulGPT
The back end makes use of these components from AWS:
• SAM
• CloudFront
• CloudFormation
• Route53
• Lambda
• API Gateway
• Secrets Manager
• S3
Education
Bachelor's Degree in Mechatronics Engineering
Sabanci University - Istanbul, Turkey
Certifications
Improving Deep Neural Networks: Hyperparameter Tuning, Regularization, and Optimization
Coursera
Neural Networks and Deep Learning
Coursera
Skills
Libraries/APIs
HealthKit, REST APIs, OpenAI API, Claude API, TensorFlow, iOS Native Libraries, Node.js
Tools
Xcode, Claude Code, Claude, Figma, ChatGPT, Firebase Cloud Messaging (FCM), Apple Vision Pro, Expo, Codex, GetStream, Amazon Elastic Container Service (ECS), ComfyUI, Whisper
Languages
Swift, Swift 5, PHP, C, ANSI C, Python, Go, TypeScript, SQL, C++, JavaScript
Frameworks
UIKit, StoreKit, iOS Frameworks, SwiftUI, Laravel, WatchKit, Core Data, Combine, CloudKit, React Native, iOS SDK, Streamlit, Fastify, LangGraph
Paradigms
REST, Mobile Development, Model Context Protocol (MCP), Microservices, Management, Agile Project Management
Platforms
iOS, Firebase, Mobile, WatchOS, Amazon Web Services (AWS), VisionOS, Docker, RevenueCat, Google Cloud Platform (GCP), Contentful, AWS Lambda, Apple TV, Algolia, RunPod
Storage
MySQL, PostgreSQL, Google Cloud, Cloud Firestore, NoSQL, Amazon S3 (AWS S3), Amazon DynamoDB, Neo4j
Other
Engineering, Software Design, Software Development, Motion Control, PLC, Human-machine Interface (HMI), Software, AVFoundation, Mobile Apps, Widgets, The Composable Architecture (TCA), Mobile UX, Mobile App Development, User Interface (UI), OpenAI, Large Language Models (LLMs), Mobile Architecture, Software Engineering, AI Tools, Product Management, Storage, Software Project Management, Concurrency, Full-stack, Architecture, Chatbots, LangChain, CTO, Artificial Intelligence (AI), FastAPI, Supabase, Retrieval-augmented Generation (RAG), AI Agents, Pgvector, Cursor AI, AI Workflow, SaaS, Security, Agentic AI, Deep Learning, Optimization, Hyperparameters, Artificial Neural Networks (ANN), Backpropagation, Neural Networks, Background Jobs, APIs, SDKs, Lottie, Identity & Access Management (IAM), Computer Vision, Machine Vision, Instant Messaging, OpenAI GPT-4 API, OpenAI GPT-3 API, Text-to-Speech (TTS), Serverless, Amazon API Gateway, Live Chat, Pinecone, Generative Pre-trained Transformers (GPT), Virtual Reality (VR), Game Development, Augmented Reality (AR), Llama 3, Amazon Bedrock, Cloudflare, Image Generation, Qdrant, Modal (Serverless GPU)
How to Work with Toptal
Toptal matches you directly with global industry experts from our network in hours—not weeks or months.
Share your needs
Choose your talent
Start your risk-free talent trial
Top talent is in high demand.
Start hiring