Summary We’re rebuilding a high-traffic AI math platform used by thousands of students each month. The site delivers step-by-step math solutions and live tutoring options. We’ve already identified key performance and accuracy issues and now need a technically strong engineer who can: Optimize our existing front-end (React/JS-based app) for speed and responsiveness. Improve solver accuracy by refining prompts, adding validation logic, and integrating reliable math libraries. Help shape the next version of our AI architecture so the app can scale to thousands of daily users and become category-leading. This role combines product engineering with AI problem-solving. Deliverables App & Performance Optimization Implement deep architectural improvements: code-splitting, lazy component loading, SSR/hydration for SEO visibility. Refactor front-end logic to allow instant transitions between solved questions (AJAX / client-side routing). Improve caching, loading indicators, and API call efficiency. Further enhance Core Web Vitals (LCP < 2.5s, TBT < 300ms). AI Solver Training & Logic Audit current solver model (OpenAI / API calls). Improve prompt engineering for consistent, step-by-step math explanations. Add mathematical validation layer (e.g. SymPy / Math.js check) before output display. Build dataset or correction loop to train and reinforce correct solutions. Experiment with hybrid models (LLM + symbolic computation). Collaboration Work with our mathematician (benchmarking competitor accuracy) to align solver improvements. Collaborate with our front-end dev (PHP) to maintain smooth 1. Have you built or optimized an AI-driven web app before? If yes, please share examples. 2. How would you approach improving solver accuracy using prompt engineering and validation logic? 3. What steps would you take to cut load time and API response delay between questions? 4. Which math or AI frameworks are you comfortable using (e.g. Math.js, SymPy, LangChain)? 5. What’s your expected timeline and estimated budget to deliver the first performance and solver improvement phase?