About

I’m seeking an experienced AI/NLP developer to build a tool that rewrites AI-generated text into natural, human-like writing and reliably passes major AI-detection systems (e.g., GPTZero, Originality.ai, Turnitin). Think of products like hix bypass, undetectable.ai, or naturalwrite.com. Key goals & success criteria : - Humanize AI-written input while preserving original meaning and intent. - Target accuracy: 95% (measured by semantic-preservation tests and human review). - The rewritten output must pass a battery of AI-detection services. I (the client) will run tests across multiple detectors; the project will only be accepted when those tests are passed. The developer must provide fixes at no extra cost if initial delivery fails detection tests (see Deliverables / Acceptance below). Core features: - Paste/upload AI-generated text → receive humanized output. - Tone/style controls (e.g., formal, conversational, concise). - Ability to handle long-form content (articles, essays). - Fast processing and scalable architecture suitable for a SaaS MVP. - Clean web UI + API endpoint (so the product can later grow into full SaaS). Deliverables: 1. Working prototype (web-based) demonstrating all core features. 2. Source code, setup & deployment documentation, and README. 3. Test report showing rewritten samples run through multiple AI detectors. 4. A short report describing the approach used to achieve detection resistance and semantic fidelity. 5. A 30-day remediation guarantee: if I test the delivered system on agreed detectors and it fails, you will fix issues at no additional cost until it passes. Acceptance criteria: 1. Functional web prototype hosted for review (staging URL). 2. Rewritten samples provided by developer must pass the specific AI detectors I use (I will provide the detector list during evaluation). 3. Semantic quality: preserved meaning with few or no factual changes, readability improved as requested. 4. All source code and docs uploaded to a repo (GitHub/Bitbucket) with clear deployment steps. Required skills & experience: - NLP, text generation, and paraphrasing expertise. - Experience with LLMs (OpenAI, HuggingFace, or custom fine-tuned models) and text-processing pipelines. - Familiarity with AI detection tools and mitigation strategies. - Full-stack web development (React/Next, Node/Python backend) and cloud deployment (AWS/GCP/Azure). - Strong testing and validation practices (automated and manual). Preferred : - Prior work on rewriting tools, plagiarism/quality tools, or text-classification systems. - Knowledge of ethical considerations and safeguards (see below). How to apply: Please include the following in your proposal: 1. Short approach: how you would build the system and why it will meet the acceptance criteria. 2. Suggested tech stack and hosting plan. 3 . Timeline and estimated cost for an MVP (staging + code). 4. Examples/links to previous relevant work (GitHub, demos, case studies). 5. Any questions you have about the detector list or acceptance tests.