Retrieval & RAG

Replace static retrieve-then-generate with autonomous agents that plan, choose sources, retrieve iteratively, reflect, and re-query.

Combine sparse lexical retrieval (BM25) with dense vector retrieval and fuse the results.

Condition the generator on top-k chunks retrieved from an external dense index so knowledge lives outside parameters.

After cheap bi-encoder or BM25 retrieval, rescore top-N candidates with a cross-encoder that jointly attends over (query, candidate).

Build an LLM-extracted entity-and-relation knowledge graph plus hierarchical community summaries, then answer global queries via map-reduce over those summaries.

Track and surface, alongside a RAG-grounded answer, which retrieved chunks supported which claims, so the binding between answer span and source survives all the way to the user.

Prepend a short LLM-generated description to each chunk before embedding so the chunk carries its situating context.

Give the agent a compact, ranked map of the codebase's symbols and their dependencies so it orients on what matters before reading any files.

Use an LLM to generate several alternative formulations of the user's query, retrieve documents for each, and rank-fuse the results so recall does not depend on one phrasing.

Build an LLM-extracted schemaless knowledge graph from the corpus and run Personalized PageRank seeded on the query's key concepts so multi-hop retrieval completes in a single pass.

Embed each query and, when its nearest cached neighbour is within a similarity threshold, return the stored answer instead of re-running the model so near-duplicate questions are answered cheaply.

Process raw documents into RAG features as a continuous stream rather than a batch job, with typed models pinning each stage.

Treat the source-of-truth document store as the only writer; keep the vector index in sync by emitting change-data-capture events onto a queue that the feature pipeline consumes.

Add a lightweight retrieval evaluator that grades each retrieved document and triggers corrective web search on poor retrievals.

Have the LLM write a hypothetical answer document, embed it, and use it as the retrieval query.

Decompose RAG into a typed three-layer hierarchy of Module Types, Modules, and Operators so the pipeline (routing, scheduling, fusion, retrieval, post-retrieval, generation) can be rearranged per que…

Train the model to ignore irrelevant retrieved documents (distractors) in a domain-specific RAG setting.

Route a query through a multi-level cascade — coarse source or index selection, then per-source narrower retrieval, then chunk-level — so each retrieval decision is pushed to the cheapest tier that c…

Fine-tune the model to emit reflection tokens that decide when to retrieve, evaluate retrieved relevance, and assess generated support.

Answer a natural-language question over a database in three stages — synthesise an executable query, run it against the data layer with model calls embedded in execution, then generate the answer fro…

Retrieve by having the model reason its way down a document's own table-of-contents tree to the relevant sections, instead of embedding chunks and ranking them by vector similarity.

Run a front-loaded phase in which an agent interviews domain experts and converts their undocumented know-how into a structured, queryable knowledge base before downstream task agents read from it.

Retrieve from a large skill library by returning each relevant skill together with its prerequisite dependency closure as an ordered subgraph, so the bundle the agent receives is executable rather th…