Looking for a job

Prachi Nikalje

@prachinikalje

Data Scientist at Paraxial Technologies building production 3D AI systems and a detection model with 94.5% mAP.

At Paraxial Technologies, I architect and deploy containerized production AI workflows for high-resolution 3D volumetric data. I engineered and validated a 3D object detection model that achieved 94.5% mAP, and streamlined inference pipelines to cut microservice response latency by 32%.

I built an AI research assistant with LangGraph and OpenAI LLMs, combining multi-agent workflows with RAG and a FastAPI backend. My work spans computer vision, NLP, and generative AI, including a speaker verification system and a text-classification competition entry that ranked 82nd out of 992 teams.

Experience

Work history, roles, and key accomplishments

Current

Data Scientist

Current

Paraxial Technologies

Nov 2025 - Present (11 months)

Architected and deployed containerized production AI workflows using Python and Docker for high-resolution 3D volumetric data, accelerating pipeline reproducibility and horizontal scalability by 25%. Engineered and validated an advanced 3D object detection model, achieving a 94.5% mean Average Precision (mAP) in identifying localized anomalous patterns within massive dense structural scans.

Python Docker MONAI NnDetection 3D Detection MLOps Microservices REST APIs

Assistant Professor

Bennett University, The Times Group

Jul 2024 - Oct 2025 (1 year 3 months)

Coached 15+ student engineering teams for national-level hackathons, steering custom AI systems and NLP software architectures from prototyping phases to cloud deployment. Formulated and delivered comprehensive graduate-level coursework in advanced Python development, Java paradigms, and enterprise software engineering structures to over 120+ students.

Python Java Software Engineering AI NLP cloud deployment Teaching