Company Description
At Cerebry, our mission is to make personalized tuition accessible to everyone, regardless of geography or family income. Our innovative AI technology serves as a personal tutor, creating unique, adaptive questions for each student based on their specific level of understanding. The AI tutor provides hints, step-by-step solutions, and can modify question difficulty to cater to individual learning needs. Our team, comprised of Ph Ds, data scientists, software engineers, and educators, is dedicated to making effective adaptive learning available to all, ensuring every student can receive the best tuition anytime, anywhere.
Role Description
As our AI Architect, you will own the end-to-end architecture and implementation of autonomous, interactive AI agents tailored for educational contexts. You’ll translate pedagogical goals into robust, cost-efficient systems that leverage state-of-the-art Large Language Models (LLMs) and Retrieval-Augmented Generation (RAG), ensuring high accuracy, minimal latency, and zero hallucinations. You will collaborate closely with product managers, curriculum designers, and cloud engineers to bring these intelligent agents into production at scale.
Qualifications
Demonstrated experience building "agentic or conversational AI systems" at scale.
Deep understanding of LLM architectures, fine-tuning methods, and prompt engineering.
Hands-on with RAG implementations, vector stores, and semantic retrieval.
Strong problem-solving abilities and attention to detail
Commitment to AI safety, ethics, and rigorous validation to prevent hallucinations and bias.
Strong grasp of algorithmic complexity, distributed systems, and cost-optimization strategies in cloud environments.
Bachelor's or Master's degree in Computer Science, Engineering, or related field
Key Responsibilities:
Architect & Build Agentic Systems: Design modular, extensible AI agents for tutoring, content creation, assessment, and adaptive learning workflows.
LLM Integration & Fine-Tuning: Select, integrate, and fine-tune models (e.g., GPT-4, LLa MA, Claude, Gemini, open-source transformers) to meet curriculum objectives and compliance standards.
Retrieval-Augmented Generation (RAG): Implement high-throughput RAG pipelines using vector databases (FAISS, Milvus, Weaviate) and secure knowledge sources to ground responses and eliminate hallucinations.
Performance & Cost Optimization: Engineer for low time complexity and cloud cost-efficiency (GCP, AWS, Azure), using caching, batching, dynamic scaling, and model distillation techniques.
Robust Validation & Safety: Develop automated testing, QA frameworks, and monitoring dashboards to detect and mitigate biases, inconsistencies, or out-of-scope outputs.
Collaboration & Mentorship: Partner with cross-functional teams—data scientists, backend engineers, UI/UX designers—to integrate agentic systems into our Ed Tech products; mentor junior engineers in best practices.
Thought Leadership: Stay abreast of breakthroughs in multi-agent systems, reinforcement learning, prompt engineering, and present findings internally; contribute to whitepapers or open-source initiatives.
