Command Palette
Search for a command to run...
Namra Maniar
Software engineer focused on backend systems and automation, with experience in security data pipelines and microservices. Skilled in containerization, observability, and CI/CD. At Optiv, working on large-scale security data ingestion, integrity monitoring, and executive-level analysis. Proficient in Python and Go, with strong system design and reliability fundamentals.
Software Engineer
@ Optiv, Bangalore
Contribute to a centralized, high-throughput Threat Management Platform for managing client findings and generating reports across multiple security assessments. Designed and developed an internal SDK to standardize platform integrations, reducing new use-case development time by 75% and enabling automation across 15+ integrations. Built ingestion service components with reliability-focused design, including event ordering, idempotent processing, and dead-letter handling to minimize data loss in high-volume systems.
SDE & AI Intern
@ Intervue.io, Bangalore
Reduced RAG pipeline latency by 50% by optimizing embedding models and tuning search patterns, improving real-time candidate analysis throughput. Built containerized microservices for a Voice AI agent (STT → LLM → TTS), designing a graph-based conversation state machine for context-aware interactions and deterministic flow control. Collaborated with system architects to refactor backend modules and data models, reducing critical production incidents by 40% through improved consistency and fault isolation.
Software Engineering Intern
@ Greenie Web, Singapore
Improved backend performance by ~25% via regex-based refactors and Postgres/MySQL query optimizations; stabilized APIs with robust error handling; automated QA by integrating Python scripts with Excel to accelerate bug triage.
Local RAG Chatbot
Built an offline retrieval-augmented-generation pipeline optimized for latency-sensitive environments. Selected HNSW approximate nearest-neighbour indexing over brute-force search, reducing retrieval time to under 50 ms at scale while keeping memory footprint low. Containerised components with Redis caching for O(1) repeated-query lookups; designed for secure, offline operation with no external API dependencies.
Version Set Manager
Developed a scalable academic submission platform with role-based access control, background job processing for large-file uploads, and JWT-based authentication. Selected MongoDB's document model over relational schemas to support flexible versioned submission records, and architected a container-friendly, horizontally scalable system for high-throughput use cases.