Open to opportunities

Bhavya Fattania

@bhavyafattania

Message

I’m an AI intern focused on responsible workflow automation and translating AI into marketing-ready adoption.

India

Message

What I'm looking for

I’m looking for an internship where I can design AI that improves marketing and workflow operations, prototype with real feedback loops, and ensure responsible adoption through clear communication for non-technical stakeholders.

I’m a Computer Science student drawn to “organizational” AI problems: mapping real handoffs, removing bottlenecks, and keeping solutions usable for the people who rely on them. At Olbuz, I observed a 6-hour, 3-person manual reporting workflow, identified where the friction lived, and prototyped an AI-powered multi-agent automation in n8n that eliminated 70% of the effort and ran without human intervention.

My larger work focused on responsible deployment: for the Government of Gujarat’s ACPC, I collaborated with officials and faculty to prototype, refine, and ship an AI assistant with explicit human-review guardrails using an OpenAI safety layer, then iterated through a beta cycle with English, Hindi, and Gujarati feedback. I also publish technical articles on Medium—breaking down LangChain/LangGraph and LLM fine-tuning—to help non-technical stakeholders adopt AI with clarity, not just capability.

Experience

Work history, roles, and key accomplishments

AI Assistant for Admissions

Government of Gujarat - ACPC

Collaborated with ACPC officials and faculty to prototype, refine, and ship an admissions AI assistant that addressed peak-season delays (hours to answers that could take ~30 seconds). Implemented responsible AI guardrails using an OpenAI 20B safety layer, ran beta feedback with stakeholders, and incorporated English, Hindi, and Gujarati.

Responsible AI Human In The Loop GuardRails OpenAI Stakeholder Management Beta Prompt Engineering Technical Documentation

AI Workflow Automation

Olbuz

Mapped a 6-hour, 3-person manual reporting workflow and prototyped an AI multi-agent automation in n8n that eliminated 70% of the effort and ran without human intervention. Iterated using an observation → bottleneck identification → prototype → feedback loop.