Machine learning is most valuable when it solves a specific business problem: predicting demand, spotting unusual transactions, sorting documents, recognising products in images or flagging defects on a production line. I focus on that practical side, starting with your data and your goal, then choosing the simplest model that reliably does the job.
I have worked on machine learning since 2019 with TensorFlow, Keras, NLP and computer vision, alongside 10+ years of Python development. That means the model doesn't stay in a notebook: I wrap it in an API, connect it to your systems and deploy it so your team can actually use it.
Forecast sales, demand, churn or prices from your historical data, with clear and explainable results.
Classify, tag and extract information from emails, reviews, tickets and documents automatically.
Face detection and recognition for attendance, access control and identity verification.
Detect and count objects, products or people in images and video streams.
OCR, image classification and quality checks that turn images into structured data.
Models packaged as APIs, deployed to the cloud and monitored so they keep performing.
Tell me about your goals and current process over WhatsApp, email or a call.
I propose the approach, tools and timeline, with a clear cost estimate before any work starts.
I build in small steps and share progress regularly, so you see results early.
I deploy, document and hand over, and stay available for fixes and improvements.
Professional Python development since 2016, with AI and machine learning work since 2019.
Most of my client work is under NDA, which is why it isn't in a public portfolio. Your project gets the same discretion.
You work with me directly, with no account managers or middlemen, in clear English and with regular updates.
From planning and development to deployment and support, one developer owns the whole solution.
“Chandni knew what I wanted and executed on that perfectly. She completed the job within a couple hours. Working with Chandni was a speedy process…”
“Ms Chandni was highly professional, friendly, quick in grasping the requirement and resolving the requirement asap.”
“Great at communicating and problem solving. Would hire her again.”
It depends on the problem. Some tasks work well with a few hundred examples, especially when starting from pre-trained models; others need more. After looking at your data I'll tell you honestly what is achievable.
Yes, and it is often the best option. Fine-tuning existing models or using modern AI APIs is usually faster and cheaper than building from zero.
I measure accuracy on data the model hasn't seen and share the results with you before deployment, so you know exactly how it performs.
Yes. Models can run on your servers, in your cloud account or on edge devices, so sensitive data never has to leave your control.
I charge $40/hr. A small proof of concept is often the best first step to test feasibility before a full build. You get an estimate after a free consultation.
Tell me about your project and I'll reply with ideas, a timeline and a quote.