Service · AI and ML
AI features and machine learning that earn their place.
YourTechConsultants, led by Raghav Palriwala in Indore, builds AI applications with the Gemini, Claude and OpenAI APIs, and machine learning and deep learning on your own data. I design each one with you, starting from the task and the records you hold. Most projects start with a written Workflow & Data Audit, so you see the plan before you commit.
Signs AI or ML can earn its place.
A task is slow because a person reads, sorts or answers text, such as messages, documents or enquiries.
You have records of past cases and a decision you want to make sooner or more consistently.
You can name the person who would check the output and act on it.
You want the right tool for the question, whether that is a model, a rule, a query or a dashboard.
What you end up with.
- Before I build
A written recommendation
It states the problem as I understand it, the data it would need and the method I would choose, and proposes how you and I would judge success.
- AI applications
A working AI application
An LLM chatbot, or a tool that answers from your own documents, built on the Gemini, Claude or OpenAI APIs.
- Machine learning
A model built for your question
A model trained on your data. The proposal would say how I would test it and how you would use the result.
Language models inside a working tool.
YourTechConsultants builds LLM chatbots on the Gemini, Claude and OpenAI APIs. It also builds retrieval-augmented generation: the model reads relevant passages from your own documents before it answers.
The WhatsApp Business CRM in the Operations & Analytics Suite uses Gemini to assist with customer messaging and campaigns. YourTechConsultants built it for Puranmal Sons. Read the Operations Suite case.
I design where a person reviews the output. Before suggesting that a feature goes live, I would try it on real questions or documents from your work.
Machine learning and deep learning.
YourTechConsultants uses scikit-learn for classical models, and PyTorch and TensorFlow for deep learning. AWS SageMaker, Athena and Glue cover data and training on AWS. NLP, which is machine learning on text, is part of the work too.
My machine-learning work spans credit-risk modelling and large-scale data pipelines (employment background below) and client projects that stay private.
How an AI or ML project runs.
- First callI ask for the task in plain words, what data exists and who would check the output.
- The AuditA feasibility audit of how well machine learning suits your data and your question, ending in a written recommendation.
The Audit ends with a written recommendation, and you can stop after it.
Already have a spec or a system? Send me your spec through the form below, or contact me at yourtechconsultants24x7@gmail.com or WhatsApp. I will say what I would do first.
Work you can check.
| AI application | Operations & Analytics Suite: a WhatsApp Business CRM that uses Gemini to assist with customer messaging and campaigns. Built by YourTechConsultants for Puranmal Sons, a wholesale building-materials distributor in Indore and a client since 2019. Its leadership commissioned, reviewed and approved the work. Read the Operations Suite case. |
|---|---|
| Background (employment, not a YTC project) | Manager, Data Scientist L3, MoneyView, 2021–24: a credit-risk model with 72%+ AUC, an SMS-data pipeline at 200 million SMS a day, and fraud-detection improvements above 450%. The same figures are on my data science page on Palrigraphy, my separate site. |
Much of YTC’s work is for private clients, so it is shown here with client details removed.
Questions about AI and ML.
How do I know whether machine learning fits my problem?
It depends on the question and on the records you hold. In the first conversation I ask about both, and the written recommendation says whether a model, a rule, a query or a dashboard fits best.
Can a chatbot answer from my own documents?
That is retrieval-augmented generation: the model reads the relevant passages before it answers. It works best on documents that are accurate and current, and I design where a person reviews the answers.
What machine-learning experience is behind this?
Credit-risk modelling and large-scale data pipelines at MoneyView (employment background, shown under Proof) and client projects that stay private. For an AI application, read the Operations Suite case, which shows a Gemini-assisted WhatsApp CRM.
Which models and tools do you use?
The Gemini, Claude and OpenAI APIs for AI applications. scikit-learn, PyTorch, TensorFlow and AWS SageMaker, Athena and Glue for machine learning. The proposal says which fits and why.
What do you need from me?
The task in plain words, a sense of what data exists, and the person who would check the output.
Tell me the task, and I will say what would fit.
I read every message and reply personally.
Write to me
Two or three sentences is plenty, for example: “My team answers the same customer questions every day” or “I have a record of past cases and want to know whether a model could sort them”. No client data, please. If you would like a call, say which days suit you.