#Dapper

AISep 21, 2026

LLMs and Transformers from First Principles — Tokenization, Attention, Embeddings, Pretraining, and Fine-tuning (with PyTorch and Hugging Face Code)

LLMs from first principles: tokenization, embeddings, RoPE, self-attention, multi-head, nanoGPT, pretraining, SFT, LoRA, QLoRA, DPO, vLLM serving.

40 min read3

AISep 19, 2026

AI Deployment and MLOps — Serving Models in Production, MLOps Best Practices, Monitoring, Cloud Strategies, and an End-to-End AI Application (with Code)

The MLOps playbook: deploying models (REST/batch/streaming), versioning + registry + CI/CD, drift monitoring, cloud strategies, and a full end-to-end app.

37 min read7

AISep 17, 2026

No Python, No PhD: Train Real ML Models in C# with ML.NET (Regression, Classification, Clustering)

How Mattrx swapped its Python ML service for ML.NET — regression, classification, and clustering trained in C#, in-process, no PhD required.

23 min read6

AISep 16, 2026

Build a RAG Chatbot in C# with Semantic Kernel and Azure AI Search in 2026 — Step by Step (Real Project, Code, Production Metrics)

Full C# RAG chatbot: Semantic Kernel + Azure AI Search hybrid + reranker + streaming + tenant security. Mattrx Help: 4% hallucination, 91% recall.

28 min read9

AISep 14, 2026

Agentic RAG vs Traditional RAG in .NET in 2026 — When Each Wins, Real Code, Production Metrics

Agentic vs Traditional RAG in .NET — when each wins, full Semantic Kernel code, the router that saved Mattrx $10,900/month on AI bills.

25 min read23

System Design

System DesignSep 13, 2026

Design YouTube — Video Upload, Transcoding, and CDN Delivery at Scale

A system design walkthrough of YouTube: resumable upload, the transcoding pipeline, blob storage, CDN delivery, and adaptive bitrate streaming at scale.

14 min read19

System Design

System DesignSep 12, 2026

Design Uber — Geospatial Matching, Live Location, and Surge Pricing

A system design walkthrough of Uber: geospatial indexing (geohash/quadtree/H3), finding nearby drivers, rider matching, live location, and surge pricing.

13 min read24

System Design

System DesignSep 11, 2026

Design a Payment System — Idempotency, Ledgers, and Exactly-Once at Scale

A system design walkthrough of a payment system: idempotency, the double-entry ledger, exactly-once processing, and reconciliation with the provider.

14 min read20

System Design

System DesignSep 9, 2026

Design WhatsApp — Real-Time Messaging, Delivery Receipts, and Presence at Scale

A system design walkthrough of WhatsApp: WebSocket delivery, message ordering, delivery and read receipts, online presence, and group fan-out at scale.

14 min read18

System Design

System DesignSep 8, 2026

Design a News Feed (Twitter/Facebook) — Fan-out, Ranking, and the Celebrity Problem

A system design walkthrough of a news feed like Twitter or Facebook: fan-out on write vs read, the celebrity problem, ranking, and caching at scale.

14 min read5

System Design

System DesignSep 5, 2026

Design a Notification System — Push, SMS & Email at Scale

A system design walkthrough of a notification system: multi-channel fan-out (push, SMS, email), queues, retries, dedup, and delivery at scale.

14 min read12

System Design

System DesignSep 4, 2026

Design a Distributed Cache — Consistent Hashing, Eviction, and Replication

A system design walkthrough of a distributed cache like Redis/Memcached: consistent hashing, eviction policies, replication, and write strategies.

14 min read11

Posts

LLMs and Transformers from First Principles — Tokenization, Attention, Embeddings, Pretraining, and Fine-tuning (with PyTorch and Hugging Face Code)

AI Deployment and MLOps — Serving Models in Production, MLOps Best Practices, Monitoring, Cloud Strategies, and an End-to-End AI Application (with Code)

No Python, No PhD: Train Real ML Models in C# with ML.NET (Regression, Classification, Clustering)

Build a RAG Chatbot in C# with Semantic Kernel and Azure AI Search in 2026 — Step by Step (Real Project, Code, Production Metrics)

Agentic RAG vs Traditional RAG in .NET in 2026 — When Each Wins, Real Code, Production Metrics

Design YouTube — Video Upload, Transcoding, and CDN Delivery at Scale

Design Uber — Geospatial Matching, Live Location, and Surge Pricing

Design a Payment System — Idempotency, Ledgers, and Exactly-Once at Scale

Design WhatsApp — Real-Time Messaging, Delivery Receipts, and Presence at Scale

Design a News Feed (Twitter/Facebook) — Fan-out, Ranking, and the Celebrity Problem

Design a Notification System — Push, SMS & Email at Scale

Design a Distributed Cache — Consistent Hashing, Eviction, and Replication

Interview questions

When should you use a service mesh (Istio / Linkerd) instead of plain Kubernetes service discovery?

What is the difference between client-side and server-side service discovery?

How does service discovery work in Kubernetes without installing Consul or Eureka?

What is the difference between liveness and readiness probes, and why does it matter for service discovery?

What is service discovery and why do microservices need it?

Why does Python dominate AI/ML development — what are the real reasons?

Tokens, context windows, and the O(n²) attention cost — what every dev should know

LLM sampling parameters — temperature, top-p, top-k — when to tune each

What is a Large Language Model (LLM) and how does it work internally?

Open-source vs proprietary LLMs — when to pick which (Llama, Mistral, GPT, Claude)

RAG vs fine-tuning — when do you choose which?

How do you chunk documents effectively for RAG?

How do you prevent hallucinations and improve answer quality in RAG?

RAG in production — security, cost, freshness, and the patterns that matter

What is RAG (Retrieval-Augmented Generation) and what problems does it solve?