Hebbian

The description states that Hebbian is an AI-powered viva coach for doctors. It provides:

• Specialty-specific interview practice stations (e.g., Core Surgical Training, Internal Medicine Training)
• Timed voice or text-based responses
• Feedback against station-specific rubrics
• Transcript review and correction
• Structured scoring with explanations
• Progress tracking across repeated attempts

The system uses OpenAI models for analysis, Firebase for authentication and storage, FastAPI for backend services, SvelteKit for frontend, and Stripe for payment processing.

Inference This is a full-stack web application designed to simulate high-stakes medical interviews and provide coaching feedback using AI-assisted rubric-based scoring. The product appears to be built as a prototype or MVP, not yet scaled for mass use.

The description states that Hebbian aims to improve interview preparation by offering:

• Realistic clinical and management stations
• Personalised practice and precise feedback
• Measurable progress tracking
• A structured alternative to static question banks or informal peer practice

It positions itself as a tool for candidates applying to competitive UK medical training programmes, using AI to deliver consistent, objective coaching.

Inference The positioning evolved from addressing a personal challenge (preparing for surgical recruitment) into a broader solution targeting a niche but high-stakes market segment. The name and concept are tied to Hebbian learning theory, suggesting an emphasis on repetition and reinforcement.

The description states that Hebbian targets doctors preparing for competitive UK medical training programmes such as:

• Core Surgical Training
• Internal Medicine Training
• Radiology
• Anaesthetics
• Higher surgical specialties

It also mentions the need for structured, objective feedback in a high-stakes environment where candidates may understand content but fail due to communication or structuring issues.

Inference The target customer is likely early-career doctors or trainees applying to highly selective UK medical residency tracks. The ICP appears to be individuals who are already engaged in formal preparation and seeking more structured, measurable feedback than traditional methods offer.

The description states that Hebbian uses Stripe for access plans and practice-credit purchases. It also mentions that candidates can choose a specialty track, respond to timed stations, and receive feedback.

However, no pricing details, subscription tiers, or monetisation strategy are provided.

Inference There is an implied freemium or pay-per-use model, likely involving credit-based access to practice sessions. However, the business model remains unconfirmed in the description.

The project was built with:

• Frontend: SvelteKit
• Backend: FastAPI
• Cloud deployment: Google Cloud Platform (GCP)
• Authentication: Firebase
• Storage: Firestore and Cloud Storage
• AI tools: OpenAI GPT models, speech-to-text API
• Payment processing: Stripe

The system supports voice practice with transcript review and correction before analysis. Feedback includes criterion-level scores, evidence from the answer, missing points, strengths, and suggested structures.

Inference The technical stack suggests a modern, scalable architecture suitable for AI-driven applications. The use of structured prompts and rubrics indicates an attempt to ensure consistent evaluation, though no data on performance or reliability is provided.

The description states that this is a self-built prototype submitted to the OpenAI 2026 hackathon. It includes:

• A working end-to-end voice practice flow
• Expert-reviewed rubrics for each station
• Feedback designed to be actionable rather than generic
• An emphasis on transparency and reproducibility

However, there is no evidence of actual users, revenue, or adoption beyond the author’s own experience.

Inference The project shows early maturity in terms of functionality and design, but lacks any traction signals such as user base, usage metrics, or commercial viability indicators.

The description does not mention direct competitors. However, it implies a gap in the market for:

• AI-powered interview coaching tailored to specific medical specialties
• Structured feedback systems that go beyond generic advice
• Tools that allow candidates to track improvement over time

It contrasts Hebbian with static question banks and informal practice methods.

Inference There is likely a small but underserved segment of the market for specialty-specific, AI-assisted interview coaching. The competitive landscape may include traditional coaching firms or general-purpose AI platforms, though none are explicitly named.

• No evidence of real-world usage or impact: The product exists only as a prototype submitted to a hackathon.
• Unproven effectiveness: There is no data on whether the feedback improves interview outcomes.
• Dependency on AI quality and consistency: The system relies heavily on OpenAI models, which may vary in performance or availability.
• Lack of scalability assumptions: No mention of how the platform would scale beyond one developer’s prototype.
• Unclear monetisation strategy: While Stripe is mentioned, no pricing model or revenue path is described.

Inference The project has strong conceptual foundations but lacks any demonstration of traction, impact, or commercial viability. It remains a proof-of-concept rather than a product in active use.

1. Has Hebbian been tested with actual doctors preparing for UK medical interviews?
2. What is the current level of accuracy and consistency in feedback generation?
3. How are the rubrics developed, reviewed, and updated?
4. Are there plans to expand beyond UK-based training programmes?
5. What is the intended pricing model and how does it align with user willingness to pay?
6. How does Hebbian handle edge cases or ambiguous answers that don’t clearly map to rubric criteria?
7. Is there any data on candidate improvement over time using the platform?