Analytics & AI Engineer · Ahmedabad, India
I'm Amisha Gupta, an analytics and AI engineer with an MBA. I build Power BI dashboards, AI assistants, automations and the data pipelines behind them, so managers spend their time making decisions instead of digging through spreadsheets.
What I build
Clear numbers with filters anyone can use. Start with the big picture, then click down to a single store, team or person to see what is going on.
For example: the SOP compliance dashboard, the field sales tracker and 26 factory dashboards.
Type a question in Microsoft Teams and get a ranked answer with the reasons spelled out. The assistant can also update your plans, but only after you confirm.
For example: the Planner Assistant and a pricing simulator that predicts sales lift.
Emails go to the right supervisor, meeting notes become tracked tasks and the weekly report drafts itself. Nobody has to remember to do it.
For example: the one-click supervisor email built with Power Automate.
Sales systems, stores, websites and ad platforms are joined into one reliable source on Azure or Snowflake. Automatic quality checks catch bad data before anyone sees a wrong figure.
For example: a unified retail data model and an automated data-quality framework.
Featured project · AI assistant
A property developer was running dozens of plans and thousands of tasks. Finding out what was late meant digging, exports and a round of calls. Now the team just asks, right inside Microsoft Teams.
Example conversation shown with sample projects, people and dates. The client's real data is kept private.
It runs inside the Microsoft 365 the team already uses. There is no new system to roll out or learn.
Acts as youIt sees only what you can see. Your access, nothing more.
Preview, then confirmEvery change is shown first. Nothing saves until you say yes.
Verified changesAfter saving, it reads Planner back to make sure what you confirmed is what happened.
Audit trailDeleting anything needs an exact confirmation and leaves a record.
Featured project · AI + Power BI + Power Automate
Built for a large restaurant chain. Machine-learning models review store camera footage and flag missed procedures. This dashboard turns those flags into scores, trends and clear next steps for every outlet and every employee, replacing manual audits.
Leaders see in five seconds how the business is doing and where to look first.
Any store opens into its own page. The Send Email button sends the supervisor a list of their staff, each person's compliance rate and pending training.
A manager types the goal, and the dashboard works out exactly how much each area has to improve. A smart-suggestion engine recommends corrective actions based on each outlet's history.
Stores are ranked on a combined score, and the business decides what matters most.
Opens with a personal greeting, then shows how this person is doing and where they need support.
Supervisors see who needs which training, next to practical tips they can use in the next shift.
Screenshots show an anonymised 30-store sample.
Point at any chart and a tooltip explains each rule in plain words, for example "Store must open at 10:00 and close at 22:00", with its compliance rate.
Customer feedback comments turn into a word cloud, so the most repeated words, such as "inconsistent" or "outstanding", are easy to spot.
Featured project · Azure + Power BI
Built for a major FMCG company whose sales reps take orders at thousands of shops every day. Data flows in automatically, and managers check it on their phones each morning: who is in the field, which shops ordered, and how this month compares with last.
How many shops were visited, how many placed an order and what sold, by category, product and rep.
Every rep's log-in time, first and last shop visit and route, plus the team's average time spent selling.
Green and red cards compare this month with the same point last month, and each rep's target sits next to what they've achieved.
Outlets are grouped as active, dormant, never visited or new, and compared with last month so lost shops get a visit before they're gone.
Employee names and the client's brand are blurred to protect confidentiality.
More projects
Four more projects for global and Indian companies. Client names are kept private.
About
I'm an analytics engineer and consultant with an MBA, so I think about return on investment, marketing and operations as much as about data models. That's why my dashboards answer the "so what?" and not only the "what".
I've delivered for FMCG, retail, healthcare and manufacturing companies, from global consumer-goods brands to restaurant chains, covering everything from the data pipeline to the screen a CEO opens in the morning.
How we'd work together
We list the questions you answer every week and the numbers you wish you had.
Excel, SQL, your ERP, POS or app. I clean it and link it into one reliable model on Azure or Snowflake.
You see working screens early, try them with real users and shape them as we go.
Alerts, emails, tasks and reports run on their own once the numbers are trusted.
Let's talk
Tell me about it. I'll show you what a dashboard, an assistant or an automation could do for your team, using your own data.
29amishagupta@gmail.com