Hello! 👋

I'm Nakshatra Rajput

AI Systems Engineer

About Me

Results-driven AI Systems Engineer with production experience architecting Agentic AI systems, high-scale RAG pipelines, and real-time Voice AI. Proven track record of reducing cloud costs by 25%, accelerating research workflows by 30%, and building systems handling 200+ concurrent calls daily. Expert in Python, FastAPI, and orchestrating LLMs with vector databases (Qdrant, FAISS, Pinecone) for complex automated reasoning.

Python FastAPI LangChain LlamaIndex PyTorch Azure AWS Docker Kubernetes Qdrant Pinecone FAISS MongoDB PostgreSQL Redis REST CI/CD

Projects

01

DRIPE: Drug Repurposing Intelligence Engine

Built an agentic drug repurposing platform using biomedical knowledge graphs, GNN-based link prediction, and RAG-driven reasoning, achieving 90% precision in drug-disease hypotheses.

LangChainRAGGNNKnowledge Graphs
02

ResFit: Science-Validated Workout Engine

Developed a B2B-ready REST API generating workout prescriptions validated against 2023-2025 research studies, achieving 95% relevance. Engineered smart caching with MongoDB TTL achieving 99.9% cache hit rate and <50ms latency.

FastAPIMongoDBRESTRedis
03

TranSys - Multimodal WebRTC Translator

Built a low-latency translation system with streaming STT/TTS pipeline, intelligent routing, and session management, achieving 40% reduction in perceived latency.

WebRTCSTT/TTSPythonReal-time Systems

Experience

AI Systems Engineer (Contract)

Freelance

Oct 2025 - Present
  • Achieved 95% retrieval accuracy by designing RAG pipelines using LangChain and Tavily for automated search and summarization of complex patent documents.
  • Reduced latency by 40% through optimization of AI infrastructure using Qdrant for high-dimensional vector search and semantic knowledge indexing.
  • Engineered modular Python services integrating LLM reasoning with external search APIs, achieving 30% reduction in research workflow time.

Software Engineer (AI)

Propeller Global Ventures

Jan 2025 - Sep 2025
  • Built real-time Voice AI agent on Azure, handling 200+ concurrent calls/day with Twilio, streaming STT/TTS, and LangChain RAG, achieving 99.9% uptime.
  • Developed Docker/Kubernetes CI/CD pipelines, reducing deployment time by 40% and achieving zero-downtime releases.
  • Optimized Azure infrastructure, reducing costs by 25% through performance dashboards, right-sizing, and prompt caching.

Research Intern (ML & Deep Learning)

Manipal University Jaipur

Jan 2024 - Jul 2024
  • Achieved 86% accuracy on X-ray severity classification using CNN models (VGG16, ResNet50) and transfer learning.
  • Improved baseline performance by 18% through hyperparameter optimization.

Education

BTech in Computer Science Engineering

Manipal University Jaipur

2020 - 2024
  • Graduated with CGPA: 8.55/10