Bajrang Jhorar

About Bajrang Jhorar

Backend Engineer with 3.5+ years of experience designing and building scalable, high-performance backend systems and production-grade GenAI / LLM-powered platforms. Fast learner with strong ownership, experienced in taking systems from architecture to production at scale.I have hands-on experience building LLM-powered RAG systems using FastAPI, Python, async I/O, Pydantic, REST APIs, and modular service architecture. I’ve designed hybrid search and retrieval pipelines combining Elasticsearch (BM25, filters, index tuning) with embedding-based vector search, improving retrieval accuracy and reducing hallucinations in real-world workloads.My core strengths include backend performance optimization, distributed system design, and data-intensive architectures. I’ve led API latency reductions of up to 90% through query optimization, indexing strategies, and async concurrency. I’ve implemented secure multi-tenant architectures with tenant-aware request context, dynamic database routing, connection pooling, and JWT-based authentication while maintaining strict p95 SLAs under peak load.On the data side, I’ve built large-scale data models using Node.js, MongoDB, Mongoose ODM, and SQL, supporting real-time aggregations across millions of records. I’ve designed fault-tolerant, asynchronous ingestion pipelines, eliminating cascading failures during high-volume data uploads.In addition to individual contribution, I’ve led a team of 6 backend engineers, delivering a multi-tenant Django / Django REST Framework platform on Microsoft Azure, driving architecture decisions, CI/CD, and production delivery.Actively Interested In• Backend / Platform Engineering roles• GenAI, LLM, RAG, Search, and AI-driven systems• Scalable, low-latency distributed systems

Experience

1. Software Engineer

   Partex

   Jul 2022 — Present · Pune, IN

   Led a team of 6 backend engineers to deliver a multi-tenant Django platform on Azure with MS SQL Server, owning architecture, performance, and establishing Docker-based deployments and DevContainer workflows for consistent development and reliable production releases.• Built production-grade LLM-powered GenAl/RAG backend using FastAPI, Pydantic, async request handling, and modular routers/services; delivered ~200-500 ms p50 and ~1.5-3.0 s p95 end-to-end response latency.• Implemented hybrid retrieval stack: Elasticsearch BM25 + metadata Filters plus embedding-based vector retrieval (chunking, top-K reranking-ready design); improved Top-K retrieval hit-rate by ~20-35% and reduced hallucination/no-answer rate by ~15-25%(standard offline relevance eval).• High-Performance Architecture: Orchestrated a search optimization strategy reducing API latency by 90% through advanced index tuning and query optimization.• Implemented secure multi-tenant backend architecture with tenant-aware request context and dynamic database routing (per-tenant credentials/DB selection), enforcing isolation at the service layer; combined SQL connection pooling (configurable pool size, reuse, timeouts) with async I/O concurrency (async endpoints + parallel execution where needed) to sustain high throughput, maintaining p95 latency under peak load.• Scalable Data Modeling: Engineered a novel hierarchical data engine (Parent/Child & Alliance logic) within Node.js and MongoDB. Successfully processed complex aggregations for of 10M+) entities in real-time without throughput degradation.• Decoupled synchronous dependencies in the bulk-ingestion module, refactoring it into a fully asynchronous, non-blocking architecture. Eliminated cascading failures during high-volume Pl uploads to ensure data integrity.• Developed a custom data versioning system for audit trails and a social-graph style notification/tagging service, enhancing user engagement and data navigability.

Education

• Indian Institute of Technology, Kanpur

  Bachelor of Technology - BTech, Civil Engineering