TRACKS
Implementation-focused deep dive into the single-node database engine substrate: workload shape, page and record layout, buffer pools, B-tree and LSM access paths, WAL, checkpoints, crash recovery, and the transaction boundary that builds on durable storage.
Architecture-level design of data platforms: source-of-truth boundaries, derived-data contracts, regional placement, serving planes, reliability controls, governance handoffs, and platform operating models.
Search Ranking And Recommendation
Index construction, lexical retrieval models, postings mechanics, ranking baselines, and the retrieval-system boundary before serving, vector search, and evaluation tracks.
Search Ranking And Recommendation
Learning to rank, experimentation, evaluation pipelines, and search quality governance.
Search Ranking And Recommendation
Crawler frontiers, politeness, canonicalization, parsing, duplicate control, freshness, corpus governance, and the handoff from web acquisition into search indexing.
Columnar storage, execution engines, warehouse architecture, vectorization, and the internals of large-scale analytical query systems.
Change capture, ingestion contracts, backfills, schema drift, and the operational trade-offs of moving data through modern pipelines.
Data models, storage trade-offs, batch versus streaming, analytical versus transactional systems, and the basic mental models for modern data stacks.
Draft track for columnar formats, table metadata layers, schema evolution, compaction, and lakehouse architecture.
Schemas, ownership, lineage graphs, discovery surfaces, and the metadata infrastructure that makes data platforms governable.
Streaming ingestion, stateful processors, watermarks, checkpoints, exactly-once claims, backpressure, replay, and the platform patterns behind low-latency data movement.
Search Ranking And Recommendation
Entity linking, graph modeling, canonicalization, and the data structures used to connect knowledge across noisy sources.
Search Ranking And Recommendation
Approximation, sketching, distributed analytics, graph mining, clustering, and operational judgment for massive datasets without mixing in personalization product loops.
Search Ranking And Recommendation
Auction design, bidding signals, marketplace objectives, and the ranking trade-offs unique to monetized retrieval systems.
Search Ranking And Recommendation
Intent parsing, reformulation, semantic matching, and the retrieval improvements that start from better representations of user needs.
Search Ranking And Recommendation
Interleaving, A/B testing, bandits, feedback loops, and the online methods used to improve ranking systems safely.
Search Ranking And Recommendation
Candidate generation, multi-stage ranking, feedback loops, experimentation, fairness, and serving architecture for personalization systems.
Search Ranking And Recommendation
Indexing basics, ranking intuition, query-document matching, and the introductory mental models behind search and recommendation quality.
Search Ranking And Recommendation
Draft track for embeddings, ANN indexes, hybrid retrieval, vector databases, and retrieval serving trade-offs.
In-memory system design through Redis as the concrete case study: event loops, data structures, persistence, replication, clustering, caching, queues, locks, and operations.
Key-value, document, wide-column, graph, and search-oriented data stores with partitioning, replication, consistency, compaction, indexing, and operations.
Search Ranking And Recommendation
Production search serving: schemas, analyzers, shards, query execution, aggregations, relevance tuning, hybrid search, indexing pipelines, cluster operations, and incidents.
Operational database depth for backend engineers: connection pools, isolation, query planning, index health, sharding, replicas, failover, and split-brain prevention.
PostgreSQL-specific depth for production systems: MVCC, WAL, locks, planner evidence, indexes, vacuum, replication, pooling, migrations, security, and operational debugging.
Single-node transaction-manager implementation: transaction state, isolation anomalies, lock managers, MVCC, optimistic validation, serializable snapshot isolation, undo, WAL integration, recovery, index consistency, testing, and observability.