Recruitment APP

4. A redundant step (saving the table)

If users exit the page without completing and submitting the form, a popup appears, adding an unnecessary step to saving their information.

Solution

1. Streamlined Job Posting: Faster, Smarter, and More Efficient

1.1. Single-Page Form vs. Stepper-Style Flow

Why I chose single-page over stepper-style:
A single-page layout allows experienced users to complete the form more quickly without page transitions. At the same time, clearly divided sections make it intuitive for first-time users to follow without getting overwhelmed.

‍Collaboration with PM/engineers or constraints:
After discussions with the PM and engineering team, we realized that a step-by-step flow would require additional logic to manage state and navigation, which would extend the development timeline. Engineers also noted that a single-page design is easier to maintain and scale.

‍Validation:
While formal A/B testing has not yet been conducted, the design choice is based on usability best practices. Research from NN/g and Baymard Institute suggests that minimizing steps and displaying key fields in a single view can enhance completion rates and reduce user friction. Post-launch, we plan to track completion rates and user feedback to further validate effectiveness.

1.2. AI Autofill + Manual Editing Option

Why I combined AI autofill with manual editing:
While AI speeds up the job posting process, its suggestions may not always align perfectly with user intent. Giving users full editing control ensures accuracy and preserves trust.

‍Technical/business considerations:
From a technical perspective, the team needed to manage draft state and transitions between AI suggestions and manual edits. However, the effort was considered manageable and worth the flexibility it offers users.

‍Decision and user validation:
We benchmarked similar products like Canva and Notion, which also combine AI suggestions with manual control.

1.3. Supporting Both AI and Manual Flows (Hybrid Approach)

Why I supported both AI and manual workflows:
Our users include experienced HR professionals who may prefer quick copy-paste methods to streamline the process of managing multiple job postings, and small business owners who may prefer job posts manually to better reflect their team culture and values. While templates offer convenience, they may not fully capture a company’s unique tone, leading some to prefer personalized descriptions. We wanted to respect their preferences rather than enforce a one-size-fits-all path.

‍Team discussion and trade-offs:
PMs were initially concerned about increased maintenance and onboarding effort. To address this, we used progressive disclosure—displaying AI suggestions first and letting users opt in or switch to manual mode—keeping the UI clean while offering flexibility.

‍Validation:
Although we haven’t conducted formal testing yet, Similar products like Workable and LinkedIn Recruiter offer AI-powered recommendations combined with manual adjustment options for resume screening. This hybrid design pattern has been proven to improve efficiency while preserving the flexibility and trust associated with human review.
‍
Moreover, according to Nielsen Norman Group’s usability heuristics, progressive disclosure helps reduce cognitive load, preventing new users from being overwhelmed by complex options.It aligns with the best practice of “showing essential information first while hiding advanced options”, which enhances task flow and improves the learnability of the interface.

1.4. Using a Shallow Step-by-Step Progress Bar to Encourage Adoption of the New Feature

Why I added a progress bar:
When we first introduced the new AI-assisted job posting feature, I anticipated that some users might feel uncertain or hesitant to try it—especially if they were unfamiliar with AI-driven tools.
To reduce that hesitation and guide them into trying the new experience, I added a three-step progress bar at the top of the form.By clearly showing that the process only involves three simple steps, the UI helps users feel that using this new feature is not complicated or overwhelming. This visual cue builds confidence and gently nudges users to proceed with the AI-assisted flow.

‍Engineering/UX constraints:
From a development standpoint, the progress bar was lightweight to implement and did not interfere with the overall layout. I designed it to be compact and visually aligned with the page structure to avoid clutter.

Decision and rationale (no testing yet):
‍Though we haven’t conducted user testing yet, research from UX experts like the Nielsen Norman Group suggests that visible progress indicators reduce uncertainty, increase perceived control, and improve task completion rates—especially for new or unfamiliar flows.

2. Improve the search experience

2.1.Smart Search & Efficient Filtering Design

Trade-off Analysis — Search Input & Filtering Experience

In this redesign of the job search interface, our primary goal was to improve search efficiency and reduce user friction. To achieve this, we implemented three key features: an intuitive input experience, intelligent keyword matching, and horizontal filtering & categorization design. Although we have not conducted formal user testing yet, our design decisions were based on common UX patterns, team discussions, and technical constraints. Below is an analysis of the key trade-offs involved:

2.1.1. Intuitive Input Experience

We introduced real-time keyword suggestions and chip-based filters to replace traditional dropdown menus or advanced filter panels. This allows users to input and adjust search conditions more fluidly and efficiently.When designing this interaction, I referenced widely adopted UI patterns from platforms like Shopee, Amazon, 104 Job Bank, and LinkedIn. These platforms commonly use autocomplete suggestions, trending keyword prompts, and chip-based filters to reduce cognitive load and enhance search speed.Therefore, I adopted a “suggestions + chips” combination to support both quick onboarding and flexible control, forming the foundation of our search bar design.

• Trade-off
  ‍We intentionally avoided complex, multi-layered filter menus to minimize user cognitive load. However, this required frontend engineering support , and maintaining clear, lightweight chip interactions. In this design, we prioritized operational efficiency and UI clarity over displaying all filters at once.
• Validation
  This design draws from proven patterns used in platforms like Shopee, Amazon, and 104 Job Bank. It aligns with Jakob Nielsen’s usability heuristics and is supported by Baymard Institute research, which confirms that such input models effectively improve search efficiency.
  Post-launch, we plan to track metrics such as search time, keyword click-through rate, and filter usage rate to evaluate performance.

2.1.2. Intelligent Keyword Matching

We implemented fuzzy search along with keyword ranking based on popularity and historical queries, aimed at reducing failed search results caused by typos or unclear phrasing.

2.1.3. Horizontal Filtering & Categorization Design

We designed a horizontally scrollable filter chip row, allowing users to quickly swipe through and combine multiple filter criteria. This approach enables more precise search results while keeping the interface clean and uncluttered.

• Why we chose real-time result count instead of a traditional search button:
  ‍Users reported frustration when they clicked the search button only to find zero or an overwhelming number of results. To reduce the need for repeated adjustments, we implemented a real-time results count that updates with each new filter. This helps users make quick decisions and reduces friction.
• PM/engineering discussion & constraints:
  ‍Engineers confirmed the feature was technically feasible without major performance impact.
• Validation:
  ‍Although no formal testing was conducted, we prioritized this feature due to its direct impact on task efficiency. This pattern is also widely used in platforms like Airbnb and Agoda. Post-launch, we plan to evaluate its effectiveness by tracking filter adjustment frequency and drop-off rates.

2.2.2. Resume Matching System

We introduced AI-powered match scores with visual indicators to help HR teams quickly identify high-potential candidates without the need to open every resume. Manual resume screening is time-consuming and prone to inconsistency. The AI-generated score helps standardize evaluation, while the visual tags indicate which specific criteria (e.g., experience, salary, location) a candidate meets—enhancing both transparency and decision speed.

Why this design over traditional manual screening?

• Manual review is costly and inefficient: According to LinkedIn Talent Trends, the average HR professional only spends 6–10 seconds per resume during the initial screening.
• AI + visual tags allow instant prioritization: HR can immediately focus on top-matching candidates without opening each profile.
• Proven in market: Platforms like LinkedIn Recruiter, Workable, and HireVue already apply similar match-based systems to improve speed and consistency while reducing bias.

PM/Engineering Discussions & Technical Constraints

To meet PM goals of reducing screening time while ensuring clarity and fairness, we designed real-time visual indicators to surface key info. To manage increased information density—especially on mobile—we used horizontal scroll with chips, keeping the UI clean. The feature is also technically feasible.

Why horizontal scroll instead of vertical stacking?

The design saves vertical space and maintains a clean candidate card layout, enabling easier side-by-side comparison—especially on mobile. It also aligns with Nielsen’s usability principles: reducing memory load, ensuring visual clarity, and providing immediate feedback.

Validation

Our approach mirrors proven systems used by LinkedIn Recruiter, Workable, and HireVue, which also rely on match scoring and structured visual cues to enhance fairness, speed, and decision quality. This confirms that our design aligns with established best practices in recruitment UX.

Future Plan validation metrics include:
Resume open rate, Average screening time per candidate, Ratio of high-match candidates making it to interview rounds, HR feedback on trust and system efficiency

2.3. Enhancing HR Efficiency:  Smart Filtering System

Trade-off Analysis — Resume Filtering

2.3.1. Quick Filtering Mechanism

To support high-volume, rapid resume screening workflows, we added features like hide viewed, bottom sheet sort menus, and persistent match score visibility, enabling HR to filter and navigate efficiently without leaving the list view.

Why this design?

We added a "hide viewed" toggle to help HR focus on unprocessed candidates, while sorting options like match score or application date are tucked into a bottom sheet to reduce clutter. Match scores remain visible across all sort types to ensure consistent comparison.

PM/Engineering Discussions & Constraints

To reduce visual overload, we used progressive disclosure—core functions are upfront, advanced filters in bottom sheets. Engineering confirmed real-time filtering is feasible and efficient.

Validation

Although we did not run formal heatmap analysis, our use of bottom sheet filters aligns with UX research from Nielsen Norman Group and Baymard Institute, which supports progressive disclosure for reducing clutter and improving discoverability.

Besides, the filtering model is aligned with best practices from tools like LinkedIn Recruiter and Greenhouse, which also use bottom sheets and persistent relevance indicators to support fast-paced HR workflows.

We plan to track the following success metrics: Usage rate of sorting options, Click-through rate on quick filter toggles, Ratio of shortlisted candidates

3. Enhanced Navigation and Visual Hierarchy

• Enhanced Navigation and Visual Hierarchy Clear Sectional Grouping: 
  The interface uses distinct sections to separate different types of information. For example, job listings, application statuses, and candidate profiles are well-structured with clear containers, reducing cognitive load. 
  Buttons like toggles, chat bubbles, and selection actions (e.g., "save" or "contact") are visually distinct and placed where users expect them, ensuring smoother interactions. 
• Improved Typography and UI Components Bold Headers & Clear Labels: 
  The use of bold and slightly larger text for important labels (such as job categories, user names, and section titles) makes scanning information effortless. 
  Contrast for Focus: 
  Different text colors and background highlights are used to indicate statuses (e.g., "in progress" buttons, job application stages). This ensures users can quickly distinguish between different states.
• Consistency Across the App Unified Design Language: 
  The layout, button styles, and iconography remain consistent across different screens, reinforcing a cohesive experience. 
  Standardized Navigation Bar: The navigation bar at the bottom ensures quick access to the most-used functions, improving familiarity and efficiency. 
• Ease of Navigation & Efficient Information Architecture Minimalist Approach: 
  Unnecessary elements are removed, allowing for a clean interface that highlights essential information. 
  Search and Filters: 
  Search bars and filtering icons make it easy for users to find relevant information without scrolling excessively. 
• Visual Differentiation of Primary vs. Secondary Functions Primary Functions (E.g., Searching, Contacting, Application Status Updates): 
  These are visually highlighted with color, icons, and strategic placement. 
  Secondary Functions (E.g., Additional Details, Saved Items):
  Presented in a subtler manner (smaller text, softer colors), ensuring they don’t distract from the main tasks.

4. Simplify the Workflow (Auto-Save Drafts for Job Posting & Profile Completion)

Trade-off Analysis — Autosave Functionality

Why we chose Autosave over manual saving
‍We implemented an autosave feature to reduce the risk of user frustration caused by accidental progress loss. Compared to traditional manual saving, autosave provides a smoother experience and allows users to focus on the content without worrying about when to click "Save."Instead of relying on users to remember to save their draft, the system saves progress automatically in the background and notifies the user with a subtle toast message. This ensures continuity and reduces cognitive load during task flow.

• Why Autosave over Manual Save?
  ‍ Improves task completion rates by removing interruption.
  Reduces the chance of lost work due to unexpected closure.
  
  Manual saving, while familiar to some users, introduces risks and delays for those who forget to save regularly.

PM / Engineering Discussions & Constraints

• PM/Engineer Feedback:
  PMs supported the feature, especially for long-form flows where it prevents data loss and user drop-off. They emphasized that using a toast message provides low-distraction feedback to build user confidence. There were no technical issues.
• Business Context:
  ‍Minimizing task abandonment and improving user trust in the platform were key business goals. Autosave directly supports both.