
Fifteen Years of Building Software for Everyday Businesses
Lessons from building business software since 2011 for distributors, clinics, restaurants, hotels and schools: customers, simplicity, support and staying power.
Artificial Intelligence
An AI and intelligent automation company built on one rule: keep human judgement in the loop.

Artificial Intelligence
AARHIT stands for Advance Automation Research & Hybrid Intelligence Technology. Rohit co-founded it on 5 October 2017 and leads its technology direction. It is a subsidiary of Acmez Technologies Pvt. Ltd. and works on AI systems that keep human judgement in the loop.
AARHIT stands for Advance Automation Research and Hybrid Intelligence Technology, and the name describes the work. It builds AI and automation systems in which software handles the predictable majority of a task and a person handles the exceptions and anything consequential.
That covers document and email handling, classification and extraction, retrieval over an organisation's own knowledge, and agentic systems whose autonomy is granted in levels rather than all at once.
Rohit co-founded AARHIT Systems on 5 October 2017 and is Co-Founder, Director and Chief Technology Officer. He leads its technology direction, which in practice means deciding where a model is appropriate, where ordinary software is better, and what review and logging every automated action needs.
AARHIT Systems is a subsidiary of Acmez Technologies Pvt. Ltd.
| Organisation | AARHIT Systems |
|---|---|
| Category | Artificial Intelligence |
| Rohit's role | Co-Founder, Director and Chief Technology Officer |
| Founded | 5 October 2017 |
| Part of | Acmez Technologies Pvt. Ltd. |
| Website | www.aarhit.com |
Insights

Lessons from building business software since 2011 for distributors, clinics, restaurants, hotels and schools: customers, simplicity, support and staying power.
Questions
Systems designed so that machine work and human judgement each do what they are good at, with the human accountable for consequential decisions.
Most work uses existing models applied carefully to an organisation's own data, with retrieval and evaluation around them. Training is considered only when a specific task cannot be solved otherwise.
Conversations
Questions about the work, collaboration or a project are welcome.