Between Bubble and Breakthrough
A reflection on the accelerating AI boom, and the hype surrounding “AI” products, and what the future might have as technology reshapes industries.
Blog
JavaScript Tips Every Beginner Should Know
A practical guide to writing better JavaScript, filled with real-world tips I’ve learned through experience.
How Image Generation Works in ChatGPT‑4o
ChatGPT-4o (GPT-4 “Omni”) is OpenAI’s multimodal version of GPT-4 released in 2024. Unlike earlier GPT models, GPT-4o can handle text, images, and audio as inputs and outputs within a single unified model.
Designing a High-Read System with Caching and Replication
Designing a system that is read-heavy, for example a product catalog API that gets 100 reads for every 1 write. Such a scenario calls for optimizations to handle a high volume of read traffic efficiently.
Indexing, Replication, and Sharding Strategies
Let’s discuss how to make databases performant and scalable. Three key techniques are indexing, Replication, and Sharding.
Data Storage & Database Systems
Choosing the right data storage is key in modern app development. In this blog, we’ll break down core concepts like relational vs. non-relational databases, data modeling, the CAP theorem, and scaling strategies such as indexing, replication, and sharding.
What if AI could teach itself to reason without our help? Hint-DeepSeek-R1
DeepSeek-R1 not only challenges the status quo but sets a new standard for reasoning performance, opening doors to a future where AI truly thinks for itself.
How Twitter Achieved Real-Time Analytics at Scale
Twitter transitioned from a Lambda Architecture to a Kappa Architecture, revolutionizing real-time event processing for 400 billion daily events
AI's Descent from the Cloud - The Next Frontier
Artificial Intelligence (AI) has traditionally relied on cloud-based infrastructures to perform complex computations. However, the technological landscape is shifting towards on-device AI processing, enabling devices to execute AI tasks locally without constant cloud connectivity.
Deliberative Alignment - Advancing Safety in AI Models
OpenAI's deliberative alignment is a training paradigm enabling AI models to explicitly reason about human-written safety specifications before responding. Used in the o-series models, it achieves high safety performance without human-labeled examples by leveraging chain-of-thought (CoT) reasoning, advancing AI safety and trustworthiness.