Shreyash G

AI Automation Consultant

Self-Employed / Freelancer • Lavoratore autonomo

Feb 2026 - Present • 8 mos

I build AI tools for small and mid-sized businesses, focused on removing manual data entry. My main product is an AI order extractor. It reads messy order emails and PDFs, pulls out product codes, quantities, prices and delivery dates, and shows them in an approval screen before export to Excel or an ERP. Built with Python, the OpenAI API and Streamlit, with checks that flag unclear values instead of guessing. I also build automated content pipelines that turn a written story into a finished captioned video using AI voice, image generation and automated editing.

AI & Automation Lead

Mindbase Education • Full time

Mar 2025 - Present • 1 yr 7 mos

I lead AI and automation for an education group, building production tools that staff use every day. What I build: - AI document processing: emails and attachments are read, classified with language models, validated and filed automatically, with errors bounced back to the sender - Automated reporting: daily performance emails with AI-written summaries, sent to leadership every morning with no manual work - Recruitment automation: CVs pulled from the hiring system, scored by AI, shortlists emailed to managers - Internal AI apps: a secure multi-model AI workspace with login and usage tracking, plus data apps and dashboards in Streamlit - Workflow automation: scheduled Power Automate flows for reminders, report distribution and data syncing across SharePoint, Excel and Outlook Previously Senior Data Analyst at the same company, promoted into this role. Tools: Python, OpenAI and Anthropic APIs, Streamlit, FastAPI, Power Automate, SharePoint, Microsoft Graph, SQL, Power BI.

Senior Data Analyst (onsite at Google)

ContinuumGlobal • Full time

Jun 2023 - Mar 2025 • 1 yr 9 mos

- Contracted through ContinuumGlobal and embedded onsite with Google's Local Guides team, the Google Maps contributor community. - Designed and analysed A/B experiments across target and holdback groups, using statistical tests (Kruskal-Wallis, Mann-Whitney U) to measure click-through, adoption and revenue impact - Built and deployed predictive models (polynomial regression, decision trees) to forecast impressions and click-through rates for homepage promotions - Led an audience analysis of India's Local Guides community, segmenting users by language, city, age and activity to explain lower engagement compared with other countries - Received the CGXI Cause of Applause award for building and deploying the Local Guides Dynamic Value Scorecard Tools: Python, SQL, Pandas, scikit-learn, Statistical testing, Dashboards.