Piyush Pandey Automation & Data Engineer
I teach machines
the boring parts.
Ten years ago I was the one typing invoices into an ERP, one at a time. Today I build the bots that do it for everyone else. This is the road between. Hop in.
Scroll to start the ride
01 2016 – 2018 Emami Frank Ross Ltd.
It began with a mountain of paper.
Data Entry Operator
Invoices. ERP screens. The same twelve keystrokes, ten thousand times a week. Most people saw a workload. I started seeing a pattern, and every pattern is a program nobody has written yet.
What it taught me: accuracy is a habit long before it's a feature.
02 2018 – 2020 Centrum Insurance Brokers
Then I learned to read the numbers.
Quality Analyst · Reporting & KPIs
KPI tracking, trend reports, and feedback loops for sales teams. I learned that a dashboard isn't a picture. It's an argument. If it doesn't change a decision on Monday morning, it's decoration.
What it taught me: data only matters when it moves someone.
03 2021 – 2022 Apollo Pharmacy
Then I ran the floor.
Store In-Charge · Data-Driven Operations
Stock, staff, revenue, and someone walking in at 11 p.m. who needs medicine now. I ran the store on MIS reports. It taught me the rule I still design by: the real user is a tired human with no time for your clever tool.
What it taught me: build for the person at the counter, not for the slide deck.
04 2022 – Now Shyam Metalics & Energy Ltd.
Now I build the machines.
Automation & Analytics Developer
I joined as an MIS Executive and became the person the plant calls when a process hurts. I build SAP GUI bots that work the night shift, ML models that solve captchas so GST lookups run on their own, scrapers, Power BI dashboards and internal web apps. Steel never sleeps, so the automation doesn't either.
- 1,500+
- hours handed back to people
- 60%
- less manual effort
- 20+
- processes rebuilt
- 24/7
- bots on shift
05 Selected work 1 / 4
Billboards from the road so far.
Intelligent Automation
Procurement Audit Automation
The problem Procurement data was pulled and checked by hand. That meant slow audits and real compliance risk.
What I built A system that extracts and validates audit data from SAP in real time, so the checks run themselves.
- Python
- SAP GUI Scripting
- SQL
Business Intelligence
Vendor Analytics Dashboard
The problem Nobody could see how vendors were really performing until something had already gone wrong.
What I built A vendor intelligence dashboard with performance tracking and anomaly detection that flags supply-chain risk early.
Data Engineering
SAP Reporting Pipeline
The problem Reports lived in a dozen places and were stitched together by hand every cycle.
What I built One pipeline that pulls, builds and distributes the reports automatically.
- Python
- SAP
- Data Warehousing
AI & Data Processing
Document Processing Engine
The problem Stacks of physical documents waiting for someone to type them in. (I've been that someone.)
What I built An OCR and LLM pipeline that reads the paper and turns it into clean, structured data.
Also on the road: ML captcha solvers for GST lookups, an invoice RPA bot that cut manual entry by 95%, and a set of internal web apps for teams that never touch SAP.