Author: Harini Kannan

I am a data scientist at Capsule8. If you wanna talk predictive models, hit me at https://twitter.com/HariniKSenthil or https://wisemachine.wordpress.com/

Quick Outline for Designing Data Pipelines for Machine Learning Projects

~ Harini Kannan ~ Leave a comment

As a Machine Learning Engineer, designing a data pipeline involves ensuring data flow is efficient, scalable, reliable, and optimized for the requirements of ML models. Here’s a structured outline to keep in mind: 1. Data Ingestion Sources and Types: Identify data sources (e.g., databases, APIs, logs, IoT devices) and data types (structured, semi-structured, unstructured). Batch … Continue reading Quick Outline for Designing Data Pipelines for Machine Learning Projects

Understanding Production RAG Systems (Retrieval Augmented Generation)

~ Harini Kannan ~ Leave a comment

1. What is RAG ? Retrieval Augmented Generation (RAG), is a method where you have a foundation model, and you have a library of personal documents – this can be unstructured data in any format. Now your goal is for answering some questions from your persona library of docs, with the help of LLM. Enter … Continue reading Understanding Production RAG Systems (Retrieval Augmented Generation)

Sample Go-To-Market Strategy for an LLM feature- Translate natural language to query language in a database company

~ Harini Kannan ~ Leave a comment

Large Language Models (LLMs) have significantly turned the tide towards AI applications being taken seriously by almost every company in every single domain. Now if you a Product Manager, you may come across the ask - Craft a Go-To-Market (GTM) strategy for an LLM feature. So here's a sample GTM strategy for a new LLM … Continue reading Sample Go-To-Market Strategy for an LLM feature- Translate natural language to query language in a database company

The Plentiful Cybersecurity Problem: Too Many Vendors, Too Little Product Sense

Featured ~ Harini Kannan ~ 1 Comment

Let’s get one thing out of the way - the market opportunity for cybersecurity as a whole - in the next 5 years, is humongous. Gartner predicts it will reach $267 billion in 2026, with an annual growth rate of 11%. If there is one domain where you find it difficult to find experienced professionals, … Continue reading The Plentiful Cybersecurity Problem: Too Many Vendors, Too Little Product Sense

Data Science Leadership Series – Part 3: Building Great Products by Balancing Product and Engineering Mindsets

Featured ~ Harini Kannan ~ 1 Comment

I wrote about why we need AI Product Management when building complex data products in a previous post in the Data Science Leadership Series, and I feel this has to be followed by a very important topic that almost every single software development team faces, especially in high growth startups where the speed of innovation … Continue reading Data Science Leadership Series – Part 3: Building Great Products by Balancing Product and Engineering Mindsets