Required stack
필요 기술
PythonAI DevelopmentMachine LearningData IntegrationSoftware ArchitectureTally ERP IntegrationREST APIODBCTensorFlowPyTorchPandas
Project brief
프로젝트 내용
I need a small AI module that plugs into my existing Tally environment, pulls raw purchase, sales and stock-movement data, and turns it into a daily report that is pushed straight back into Tally as a ledger or voucher entry I can open like any other report.
The module has to do three concrete things for every run:
• Calculate current inventory levels for each SKU in real time.
• Work out key supply-chain metrics such as reorder point, days of cover and in-transit quantities.
• Produce an accurate stock valuation that respects my weighted-average method already configured in Tally.
I do not mind whether you build the model in TensorFlow, PyTorch or another modern framework—use whatever lets you finish quickly and keep maintenance simple. A neat Python script with an .exe wrapper, or a light web service that Tally can call through ODBC or REST, would both be fine as long as the end result lands in Tally every day without manual intervention.
Acceptance criteria
• A single-click (or scheduled) job that fetches data from Tally, runs the calculations and writes the finished report back.
• Clear mapping documentation so I can see which Tally fields feed each metric.
• Source code, brief setup guide and a quick Loom/Zoom walkthrough showing the process working on my machine.
If you can demonstrate something similar you’ve automated for accounting software before, that will speed up my decision.
The module has to do three concrete things for every run:
• Calculate current inventory levels for each SKU in real time.
• Work out key supply-chain metrics such as reorder point, days of cover and in-transit quantities.
• Produce an accurate stock valuation that respects my weighted-average method already configured in Tally.
I do not mind whether you build the model in TensorFlow, PyTorch or another modern framework—use whatever lets you finish quickly and keep maintenance simple. A neat Python script with an .exe wrapper, or a light web service that Tally can call through ODBC or REST, would both be fine as long as the end result lands in Tally every day without manual intervention.
Acceptance criteria
• A single-click (or scheduled) job that fetches data from Tally, runs the calculations and writes the finished report back.
• Clear mapping documentation so I can see which Tally fields feed each metric.
• Source code, brief setup guide and a quick Loom/Zoom walkthrough showing the process working on my machine.
If you can demonstrate something similar you’ve automated for accounting software before, that will speed up my decision.