Christian C.

Key Skills

Software

Other

Don't disclose

Top Subject Matter

Image segmentation for browser-based interactive annotation and mask generation

Clinic workflow software and offline data robustness (AI agent implemented code)

AI tutoring/debate coaching platform (transcription + AI responses)

Top Data Types

Image

Text

Document

Top Task Types

Segmentation

Text Generation

Transcription

Bounding Box

Entity (NER) Classification

Freelancer Overview

Image Highlighter & Object Segmentation App — System Architect & Designer. Brings 3+ years of professional experience across complex professional workflows, research, and quality-focused execution. Core strengths include Other, Internal, and Proprietary Tooling. Education includes Bachelor of Science, Jomo Kenyatta University of Agriculture & Technology (JKUAT) (2025) and KCSE Certificate, Alliance High School (2024). AI-training focus includes data types such as Image, Medical, and DICOM and labeling workflows including Segmentation, Computer Programming, and Coding.

Labeling Experience

AndikaFlash & Andika — Product Visionary & System Architect (Andika)

Don't discloseText

Directed development of an AI gradebook that provides instant rubric-compliant grading and feedback for handwritten English and Kiswahili compositions. Used Gemini Flash multimodal OCR to interpret handwritten submissions, then generated feedback and analytics using Firestore. Specified database schemas and UX/edge-case handling while directing AI coding agents through end-to-end resolution. • Multimodal OCR of handwritten compositions via Gemini Flash • Rubric-compliant grading and feedback generation workflow • Firestore analytics and Firebase Auth/Cloud Functions integration • Edge-case handling for OCR inputs and network/timeouts

2025 - Present

AndikaFlash & Andika — Product Visionary & System Architect (AndikaFlash)

Don't discloseImageSegmentation

Directed development of handwriting recognition and conversion features that transform images of handwritten notes into structured outputs. Orchestrated AI coding agents using Gemini Vision OCR to extract handwritten content and generate PDF or editable Word documents. Defined UX flows and edge-case handling for OCR-related issues such as network timeouts and image downscaling. • Photo-to-PDF and photo-to-Word conversion using Gemini Vision OCR • Firebase integration for storage/auth as part of the app backend • Defined schemas, UX flows, and OCR edge-case resolution • Google Play Billing integrated for monetization-ready release

2025 - Present

ispeak — Founder (In Development)

TextText Generation

Directed the creation of an AI debate and public speaking practice platform that uses speech-to-text and multimodal/LLM generation components. Orchestrated AI coding agents to build the application ahead of a Gumroad launch, including defining the system architecture and verification expectations. Focused on end-to-end product readiness by coordinating tooling selection and implementation planning through the release workflow. • Platform built with Whisper for transcription and Gemini 2.5 Flash for AI generation • ElevenLabs used for voice/speaking-related capabilities • Next.js 14 and Firebase used for frontend/backend integration • AI coding agent direction for full build and launch preparation

2025 - Present

AfyaCare Clinic Portal — Product Manager & System Architect

Led development of an offline-first clinic portal where AI coding agents implemented 100% of the code under his product and system direction. Identified and mitigated data-loss risks in patient check-in by directing refactoring of the patient registry into a hybrid tabbed workspace model. Directed multiple robustness fixes affecting offline serialization, window-resize layout stability, and dashboard density for improved clinical workflow efficiency. • Offline-first desktop health centre management system implementation • Registry refactor to prevent draft data loss across concurrent consults • Directed Windows runner changes to prevent resize layout crashes • Directed offline JSON serialization handling for Firestore timestamps/datetimes

2025 - Present

Image Highlighter & Object Segmentation App — System Architect & Designer

OtherImageSegmentation

Designed a client-side image segmentation workflow using SlimSAM to generate object masks from user interactions in a browser UI. Directed the implementation of a click-to-segment “magic wand” feature that triggers boundary decoding and renders mask overlays. Specified a responsive architecture by offloading embedding/decoding work to a Web Worker to keep the UI responsive during inference. • Client-side inference in Transformers.js with no server/GPU dependency • React-Konva canvas mask overlay rendering and annotation modes • Manual bounding box annotation support alongside automatic masking • Multithreaded Web Worker design for inference responsiveness

2025 - Present

Show more

Education

Work History