· Created an efficient and mentally easy process for users to establish their professional profiles.
· Identified the connection point between Hiintern and GenAI.
· Demonstrated the technical feasibility of developing AI-powered features.
Hiintern is an internship platform. It focuses on connecting students with optimal internship opportunities, and assists employers in efficiently finding their ideal interns.
Hiintern made its debut in February 2024, while everyone’s attention was caught by AI-related products. The debut was silent, with only a few new users attracted.
As more and more companies integrate AI into their products, the company leader wanted to release an AI feature for students at minimal cost as soon as possible.
Since no one in this team had AI-related working experience or basic knowledge, a bunch of researches were conducted to obtain general knowledge about AI and OpenAI API.
Powerful content generation capabilities
Extensive general knowledge of human resource topics
Proficiency in natural language analysis
High API costs
Character length limits
Hallucinations caused by improper prompts
Even though we gained a few users several months after the debut, we still identified some essential problems from the data we collected. The completion of their profiles is not ideal.
of the profile completion on our platform is under 80% (the required completion rate to apply for jobs on Hiintern).
of new users complete the entire onboarding process without skipping the steps.
I conducted interviews with 4 users to uncover deeper reasons of low profile completion rate, and understand their attitudes towards seeking internships with AI assistance.
After the research, I identified a connection point between GenAI and students’ internship-seeking process — using AI to help them craft personal profiles. This could be the most actionable approach for both the users and the product provider.
Integrate AI into students internship seeking process with the minimal cost
Increase investor appeal
Leverage AI capabilities to assist users in creating their profiles
Ensure seamless integration of the new AI features with the current product
In the current onboarding process, without AI assistance, we help users craft their basic profile foundation. The AI can become involved in this phase to assist users in creating a more complete profile efficiently and easily.
Users experienced frustration while filling out forms and requested writing guidance during this phase, where AI can assist by leveraging its extensive HR-related knowledge.
With AI assistance, users could make clearer judgments about what to include in their profiles to align with recruiters’ requirements, helping to reduce mental strain.
Even though users requested full control during the user interviews, it wasn’t entirely feasible to implement this in the first ideation and prototype. However, the user testing provided me with a deeper understanding of user control and their expected interactions.
“What if I’m not satisfied with the AI-generated content, since the previous texts will be automatically replaced after I click the Generate button?”
“Sometimes the AI generation may be too long, and I don’t think it’s a good idea to place it above the job description on mobile device, as it may lead to confusion and make people think they are the same thing.”
Let users decide if they want to add the generations into their profile forms.
I collaborated with the backend team and conducted prompt experiments to see if the OpenAI API could provide relevant and accurate outputs based on the given inputs. The experiments felt like a kind of magical modulation.
...
# Define work experience details
job_title = """Software Engineer"""
company = """ABC Tech"""
location = """San Francisco, CA"""
start_date = """January 2021"""
end_date = """Present"""
is_internship = False
# The prompt
prompt = f"""
Write a concise work experience description within 500 characters for the role of {job_title} at {company},
located in {location}, from {start_date} to {end_date}."""
response = get_completion(prompt)
print(response)
Unfortunately, the company went out of business due to some investment issues. However, this project introduced me to AI design and opened up new possibilities.
When I reflect on this project, I still see areas for improvement. For instance, we could have provided more AI-driven suggestions in the internship search process—not just on how to refine profiles, but also offering insights about companies, which many users care about.
Integrating AI into current products is inevitable, and I will continue to focus on this area.
Hi, I’m Zoe Zhou! 👋 A UX/UI designer with 3 years of experience in web and mobile design. I’m currently seeking a full-time UX/UI position where I can bring user-friendly, clear, and delightful designs to life. Let’s collaborate to make your projects shine!
