EmpowHER.
Practical tools for the next career step.
An AI-powered career-development platform connecting interview practice, resume analysis, learning resources, and event discovery.
Team hackathon project · AI responses and résumé-analysis integrations
Fidelity FidHacks 2025 / Four-tool career platform
A team-built hackathon project. Ayug’s supplied experience describes AI response and résumé-analysis integrations. The case study covers the complete product while distinguishing the public frontend from integrations whose backend code is not present.
Inside this case file 08 sections
Product Overview
EmpowHER combines interview practice, résumé feedback, learning resources, and event discovery in a career-development experience. The project won first place at Fidelity FidHacks 2025, an invite-only event with 48 selected participants.
The platform’s product scope is broader than a chatbot: someone can practice an interview, inspect résumé feedback, find a learning resource, and discover a career event through related pages in the same experience.
- CONVERSATIONInterview practice
Guided prompts + chat API request
- FEEDBACKRésumé analysis
Text upload + local scoring criteria
- RESOURCESLearning resources
Curated career and learning links
- OPPORTUNITYEvent discovery
Preset events + browser-side additions
Shared HTML/CSS + JavaScript navigation
Chat expects a local API; its backend is not in the public repository
Original product architecture illustration. The résumé scorer is deterministic browser logic in the checked-in version.
Problem and Solution
Interview preparation, résumé improvement, resource discovery, and professional events are often treated as separate tasks. EmpowHER brings them into a common entry point with a supportive career-development voice.
The hackathon build prioritizes concrete activities instead of requiring a complex account or infrastructure setup. The resulting pages make each tool directly accessible while keeping a shared navigation and visual language.
The Complete Toolkit
Each part of the frontend supports a different preparation task. The inspected code also makes clear where a feature is driven by local browser logic and where it expects an external response.
Interview practice
A conversational HERa interface with a guided mock-interview question sequence and a separate chat request path.
Résumé analysis and scoring
Upload a text résumé, select a target role, and receive a score and feedback based on contact information, structure, action verbs, measurable results, role keywords, and length.
Career resources
A curated set of learning, networking, communication, and confidence-building links arranged for direct browsing.
Event discovery
A browser-rendered event collection with preset entries and a form for adding an event to the current experience.
Application Architecture
The public source is a multi-page HTML/CSS and JavaScript application served locally with Servor. Shared navigation connects the chat, résumé, resources, and event pages. Feature logic lives in inline scripts and a separate résumé-scoring script.
Résumé input is read with FileReader and evaluated locally. Event entries are rendered from client-side data. The chat page posts a message to a local /chat endpoint. The corresponding AI backend implementation is not included in the checked-in repository, so its model, hosting, and persistence are not asserted here.
No implemented production database or server-side authentication system was verified in this source. Local browser state is used in navigation/account affordances; it should not be described as secure identity management.
AI and API Integrations
The chat UI supports sending a message and displaying an API response, with a local endpoint contract for an AI mentor interaction. Guided interview mode uses a predefined sequence of prompts in the browser.
The checked-in résumé scorer is deterministic JavaScript: it reads a text file and evaluates explicit criteria and role-keyword matches. It is not an LLM-based résumé-scoring model in this public version. The broader project was presented as an AI-powered career platform; this portfolio makes that implementation distinction explicit.
The resource page contains curated outbound links, and the events page uses browser data. Neither is represented here as a live third-party event feed or a search integration that the source does not show.
Reliability and Scope
The résumé flow checks for an uploaded .txt file and reports a user-facing error when input is missing or incompatible. The chat interface includes an error/fallback path around its network request.
No automated test suite is configured in the inspected package manifest. The portfolio therefore does not claim test coverage, validated AI accuracy, production security, or persistent event storage. Those would be the next foundations for turning the hackathon concept into a durable service.
Challenges & solutions
The engineering challenge was connecting four career-preparation activities in a limited build window. A lightweight multi-page structure let the team focus on the core interactions and link them into a complete demonstration.
Several feature types
A shared navigation ties together conversational practice, file-based scoring, curated content, and event browsing.
Feedback that is easy to inspect
The résumé scorer exposes concrete criteria and role keywords rather than presenting its number as a validated hiring prediction.
An integration boundary to improve
A deployable, protected AI service, reproducible setup, and automated workflow tests would make the chat contract more reliable beyond the event demonstration.
Outcome and Lessons
EmpowHER earned first place at the invite-only Fidelity FidHacks 2025 event. Its four-feature scope demonstrated a connected career-development concept rather than a single isolated AI interaction.
The lasting engineering lesson is to distinguish a compelling product workflow from the guarantees of an early implementation. Clear interfaces, inspectable scoring logic, and honest integration boundaries are the starting point for a more reliable next version.
Evidence & scope
Reviewed the public repository and matching local project files: chat page, résumé scorer, browser scoring logic, resources, events, navigation, and package configuration. The event page confirms 48 participants; the first-place result is supplied by Ayug.