From theory to the systems that demand real understanding.
Long-form pieces on how production-AI teams orchestrate agents, serve models, and build the stack, plus GATE DA essays that turn probability, linear algebra, DBMS, ML, and AI into durable concepts.
LLM recommenders: Netflix, Spotify, and the post-collaborative-filtering era
Recommendation systems are absorbing LLMs, but almost never the way Twitter expects. Netflix and Spotify both run hybrid stacks — classical CF and content-based ranking as the workhorse, LLMs at the edges for explanation, cold start, and conversation. Pure LLM-as-recommender doesn't survive contact with YouTube-scale latency or freshness.
Multi-hop RAG without the hallucinations
Single-hop RAG works until your users start asking questions that require synthesis. Naive multi-hop hallucinates intermediate facts. Here's the working architecture — explicit decomposition, iterative retrieval, self-verification — with the production examples from Harvey and Hebbia that actually ship it.
GraphRAG: when knowledge graphs beat vector search
Microsoft's GraphRAG paper showed that building a knowledge graph from documents and querying community summaries beats vector RAG on global questions. Two years on, the receipts are mixed — the technique works, but the indexing bill is what most teams remember.
Agentic RAG: when the agent decides whether to retrieve
Naive RAG retrieves on every query. Agentic RAG turns retrieval into a tool the agent calls when the answer isn't already obvious — and the production wins are about latency and accuracy at once, not either alone.
Long-context vs RAG: the 2026 verdict
Gemini's 2M-token context, GPT-5's million, Claude's 200K. The argument that long context would kill RAG turned out wrong — but for reasons more interesting than 'cost is high'. Here's where the equilibrium actually settled.
PDF parsing remains unsolved: LlamaParse, Reducto, Unstructured, Marker
Two years into the production RAG era, the single biggest blocker for most enterprise deployments isn't the LLM, the vector store, or the retrieval algorithm. It's whether your PDF parser got the table right. The vendors have multiplied; the problem hasn't been solved.
RAG evaluation in production: Ragas, Phoenix, Braintrust
RAG evaluation is harder than LLM evaluation. Retrieval and generation fail for different reasons and need different metrics. Two years after Ragas formalized the four-metric core, the tooling has split into three camps — here's what each one actually buys you.
Fine-tuning vs RAG: the settled debate of 2026
Three years of Twitter brawling about whether to fine-tune or retrieve has ended in a boring, useful answer: RAG for knowledge, fine-tuning for behaviour, and both when you actually need both. Here's the decision table production teams actually use.
KV cache management: paged attention, prefix caching, LMCache
The KV cache is the dominant memory bottleneck in LLM serving, and three ideas — PagedAttention, prefix caching, and cross-instance LMCache — have rewired how it's managed. Here's how each layer earns its place in production.
NVIDIA Dynamo, vLLM, SGLang: serving stacks at scale
NVIDIA's Dynamo (GTC 2025) reframes LLM serving around prefill/decode disaggregation. vLLM, SGLang and TensorRT-LLM all sit underneath. Here's how the four-layer stack actually works, what the throughput numbers really mean, and who picks which.
Inference routing: sending each query to the cheapest model that can answer it
By 2026 the biggest lever on inference cost isn't quantisation or batching — it's deciding which model touches each query. Four routing patterns, three rounds of vendor consolidation, and a real case study where a customer support agent cut spend 80% with no measurable quality drop.
Serverless LLM platforms: Modal, Together AI, Fireworks, Replicate
Four platforms, four different bets on what serverless inference should be. Modal is the GPU-as-Python-function. Together is the model marketplace. Fireworks is the speed obsessive. Replicate is the prosumer API. Here's who picks which, and the cost numbers behind each.
Structured outputs engineering: JSON mode, function calling, constrained decoding
Three families of techniques get LLMs to return parseable data. One of them is a guarantee. The other two are negotiations. Here's when each one earns its keep — and the production regression that nobody warns you about.
CrewAI vs LangGraph vs AutoGen: the ecosystem reality of 2026
Three years into the agent framework era, only two of those three names belong in the same sentence in 2026. Here's the actual usage picture — stars, downloads, who's running what in production, and where each one quietly wins.
DSPy: declarative prompting in production
Stanford's DSPy proposes 'programming, not prompting' — declare a signature, compose modules, let an optimizer write the prompt. The pitch is irresistible; the production reality is more selective. Where DSPy actually ships, and why most teams quietly keep their prompts.
Pydantic AI: typed agents for the Python ecosystem
The Pydantic team built the framework most FastAPI engineers were quietly wishing for. Schema-first, model-agnostic, async-native — the agent framework that feels like Python production code instead of a research demo.
Inspect AI: the UK AISI's eval framework everyone copied
A government safety institute released an MIT-licensed eval framework in 2024. By 2026 it's the de facto standard at Anthropic, OpenAI, DeepMind, and the enterprise teams that take evals seriously. Here's why it won — and what the alternatives still get right.