Anurag Mishra

I’m a Senior Software Engineer focused on building high-throughput search, indexing, and AI-powered matching microservices at scale. I’ve led end-to-end delivery of hybrid retrieval pipelines that fuse KNN vector search with lexical ranking for consistently better results.

At InfoEdge India Ltd., I engineered a Job-Candidate Matching Engine combining 4 parallel KNN streams with BM25 via Reciprocal Rank Fusion (RRF). By diagnosing Elasticsearch connection-pool exhaustion, switching to async parallel execution (CompletableFuture), and adding TTL-based in-memory caching, I reduced matcher latency from 24s to ~150ms (99% improvement).

I also built ML-driven feed personalization and a CV skill-match API integrated with CV parsing outputs to achieve ~93% accuracy on live recruiter data. Beyond ranking, I delivered Similar-Jobs/Similar-Candidates features with paginated scored retrieval and caching for repeat query speed.

Earlier at Decimal Technologies, I automated AWS deployments (83% faster) and stabilized a large Elasticsearch cluster through mapping fixes, shard/memory leak remediation, and performance tuning. I’m energized by observability-driven engineering, distributed-system reliability, and building systems that stay fast under real-world load.