From Student Project To Real AI Work: Harsh Bhanushali And Dualmind Lab!

While many students learn technology in theory, Harsh Bhanushali Jitendrabhai is learning it by building and shipping real products. A second-year B.Tech Computer Science student (Diploma-to-Degree) at Parul University, he runs his own AI initiatives under Dualmind Lab and has already worked with international clients. His story is a template for student-led technology.

Hailing from the heart of Gujarat, Kutch, and raised in Rajkot, Harsh scored 93% in 10th class, and then he completed his diploma at Parul University with an 8.93 CGPA. He chose a diploma, as he was more driven by computer science and development. He was even more drawn towards building websites and creating techno-creative solutions that matter. As AI advanced, he started using AI tools as a learning aid, experimenting his way from general development toward generative AI and automation. His method is simple and repeatable: when something interests him, he digs in, researches it, and builds with it.

Having already experienced university life during his diploma, Harsh chose Parul University for its practical, project-friendly environment, its placement opportunities, and its student support; a place to keep building his skills while earning his degree.

Perhaps the most remarkable part of Harsh’s story is his freelancing. Through cold outreach, sending hundreds of emails to potential clients, he secured a 200-hour contract worth $8,000 with a UK web-development agency wanting to add AI automation to its workflow and later a 100-hour, $4,000 contract with an Australian client. He has openly stated that it was an arduous experience: he submitted approximately 700-800 emails for his first opportunity and only received a small number of responses. Additionally, he had to educate himself on email reputation and how to pace out his submissions. However, persistence prevailed. He also made sure he didn’t have all the answers prior to beginning; as the project evolved, he needed to learn the appropriate technology in order to complete the tasks at hand. He found this to be one of the many lessons of doing actual work: you will always be continuing your education versus having complete knowledge of what you are attempting when you begin.

Harsh runs his projects under Dualmind Lab, an umbrella for his technology work rather than a single product. Under it, he has built AI automation systems, voice agents, web applications, and AI infrastructure. His freelancing sharpened one skill in particular: identifying repetitive human tasks and automating them with AI, while keeping a “human-in-the-loop” where human judgment still matters.

One project shows how far AI has spread beyond tech companies. For a salon in Jamnagar, Harsh built a voice-agent solution and managed the business’s wider digital presence, including its website. It taught him how AI can serve businesses that are not themselves technology companies, exactly the kind of practical, real-world application that makes a student’s skills valuable.

A flagship project under the Dualmind ecosystem is Dualmind Arena, a platform for comparing AI models. With so many models available, ordinary users rarely know which is best for a given task. Harsh’s idea is elegant: show users responses from different models without revealing which is which, let them vote on the better answer, and then reveal the identities, building a benchmark driven by real human judgement. The first version has already drawn more than 300 users, including from the United States and Canada, and he is building a second version with early access open.

Building at this level brought both help and hurdles. Harsh earned support from startup programs, receiving cloud credits from Cloudflare’s startup program (after moving his hosting there and applying), along with support from Sarvam and Cloudcar. But he also faced real challenges: as his platform grew, so did the cost of running AI services, and at one point his site was hit with millions of malicious requests in a short span, spiking his AI costs. He responded by hardening the site with security measures, including a Web Application Firewall (WAF). The lesson stuck: building a product is far more than writing code; security, infrastructure, traffic management, and cost control matter just as much.

The biggest benefit of running Dualmind Lab, Harsh says, is the sheer number of real problems he meets daily, from staging and merge conflicts to development workflows, automation tools like Make.com, and project management. AI can assist with coding, he notes, but developers still have to understand the problem, state the requirements correctly, and decide how a solution should be implemented. Problem-solving and consistency, more than any single tool, have become the core of his growth.

Asked what he would tell other Parul University students who have ideas but hesitate because they are “just students,” his answer is a single word.

Start.
– Harsh Bhanushali, on beginning before you feel ready

His own trajectory demonstrates this: he started experimenting as a diploma student, continued to develop during his studies, and ultimately thrived from studying to real international contracts. He has a balanced perspective on startup versus placement, where he believes that a startup is a long-term opportunity but sometimes may not provide immediate stability, and thus placement and entrepreneurship are complementary rather than mutually exclusive. That’s the wisdom to borrow for a student builder. With the right blend of entrepreneurship, you can turn your AI interest into a career by pursuing a B.Tech CSE in Artificial Intelligence, a B.Tech in Artificial Intelligence – ML & Robotics, and a B.Tech in Artificial Intelligence and Data Science.