NR

Nilaa Raghunathan

Available to hire

I am an AI engineer with experience building production-ready generative AI systems that move from raw user input to reliable, real-world outputs.

About

I am an AI engineer with experience building production-ready generative AI systems that move from raw user input to reliable, real-world outputs. My work spans prompt engineering, workflow design, and evaluation of AI models for consistency and quality, with a strong focus on practical constraints rather than demos. I have built and deployed AI systems for content generation, recommendation, and decision support, combining careful prompt design with validation and iteration. What sets me apart is my attention to detail and my ability to translate abstract requirements into repeatable, high-quality AI workflows that are ready for real use, not just experimentation.

  • 1 year's experience
  • Joined February 2026

Work experience

Machine Learning Intern · Limex TradingSep 2025 – Dec 2025 · 4 mos

Led the design and validation of a portfolio risk forecasting framework using factor-based and covariance models (PFA, GARCH variants, DCC, Ledoit-Wolf) across 102 assets, boosting forecasting accuracy by 29%. Implemented rolling-window experiments and backtesting across 71 model configurations to detect under-forecasting in low-volatility regimes (78%). Established a Monte Carlo-based evaluation framework to validate tail risk (VFR=1.08) in high-volatility regimes.

Data Science Intern · Magna International Inc.Jun 2025 – Aug 2025 · 3 mos

Enhanced an ML-based inventory demand forecasting system, reducing safety stock by 12% and saving $1.4M annually. Increased forecast accuracy by 18% (MAPE) using probabilistic time-series modeling (GMM-HMM). Built a context-aware NLP pipeline (BiLSTM-RoBERTa) over 7K+ supplier records. Lowered stock-out risk by 13% via ANOVA, A/B testing & causal impact analysis. Delivered a Streamlit dashboard on Databricks with a MySQL backend to support Monte Carlo simulations for cross-functional teams.

Applied Research Intern · Vellore Institute of TechnologyApr 2022 – Jul 2023 · 1 yr 4 mos

Published first-author IEEE Access paper on sentiment analysis deployment, identifying multilingual, cross-domain, and small-scale failure modes across transformer models to establish rigorous ML evaluation practices. Evaluated mitigation strategies (transfer learning, multilingual embeddings, modality fusion) to improve scalability and reliability.

Education

Master of Science in Data Science · Columbia University2024 – 2025
Bachelor of Technology in Computer Science with Specialization in Data Science · Vellore Institute of Technology2020 – 2024

Hire Nilaa Raghunathan today

Post a job and invite Nilaa, or contact them directly. No hidden commission.

Post a job