Aditya Sikhwal

Staff Data Scientist with 6+ years of global experience applying statistical modelling and machine learning — from classical econometric methods (Logistic Regression, Random Forest, Gradient Boosting, SVM, KNN, K-Means) to deep learning and NLP/LLM systems — to solve high-stakes business problems in fintech, banking risk, and fleet-safety technology. Built and validated credit risk and anomaly/outlier detection models within regulated financial environments (J.P.

Morgan

Chase, CCAR), and applied classification techniques to a credit card fraud detection problem, giving direct exposure to the model governance, class-imbalance handling, and regulatory rigor central to financial-crime and fraud analytics.

Strong

Python programmer (object-oriented design, package development, Git) with production experience in Apache Spark, Scikit-learn, PyTorch, and TensorFlow. Proven mentor of junior data scientists, clear communicator of complex technical concepts to non-technical stakeholders, and a self-starter who partners across Product, Engineering, and Finance — including cross-functional and remote/distributed teams — to keep organizations moving toward being genuinely data-driven.

Staff Data Scientist - Netradyne Technology India Pvt Ltd

(2025-02)

• Improved a real-time accident-prediction (outlier/rare-event classification) model for commercial fleets through advanced feature engineering and signal filtering on IMU sensor data, then Gradient Boosting (LGBM) model training — raising precision from 85% to 95% and recall from 85% to 90%.
• Trained a multimodal neural network (LSTM combined with YOLOv5 vision features) to improve minor-accident prediction precision from 20% to 45%, matching model architecture to problem structure across classical and deep-learning approaches.
• Designed and shipped a RAG-based LLM chatbot with vector search so fleet safety managers can query product documentation, driver records, coaching history, and device status in natural language.
• Extended the RAG/agentic pipeline with tool-calling capabilities for on-demand driver-performance ranking and trend visualization across configurable time windows.
• Mentored and coached a team of 4 junior data scientists on experimentation design, causal inference, and stakeholder communication, embedding a collaborative, semi-remote way of working.

Senior Data Scientist - Netradyne Technology India Pvt Ltd

(2024-01 - 2025-01)

• Built anomaly/outlier detection models for IoT device health, cutting false alerts by 37% and improving real-time customer experience and device longevity.
• Led A/B testing and experimentation programs that lifted trial-to-paid conversion by 22%, translating statistical results into a business narrative that drove a $200K revenue increase.

Data Scientist - Netradyne Technology India Pvt Ltd

(2022-06 - 2023-12)

• Developed behaviour-based scoring and cohort clustering/retention models (unsupervised + classification techniques) that boosted user retention by 15%, partnering with Product and Engineering on metric instrumentation.
• Built a driving-behaviour recommendation engine using a Gaussian Naive Bayes classifier for interpretability on 6 weeks of sensor data; the resulting risky-driving alerts identified 45% of accident-prone drivers within 2 weeks.

Senior Engineer - Samsung Semiconductor India Research

(2021-06 - 2022-06)

• Improved 5G uplink throughput via advanced statistical noise estimation and model-based signal processing.
• Built deep learning models for channel parameter estimation, improving network efficiency.

Quantitative Analyst – Credit Risk (Banking & Finance) - J.P. Morgan Chase

(2020-07 - 2021-06)

• Led Python-based quantitative financial modelling for CCAR 2020 stress testing, forecasting credit losses, revenue, and risk-weighted assets across portfolios in a production regulatory environment — direct experience applying data science to financial risk detection and regulatory compliance.
• Developed and independently validated credit risk models to ensure compliance with regulatory frameworks; performed risk-adjusted profitability analysis supporting capital allocation decisions.
• Translated complex model outputs into clear business narratives for senior, non-technical stakeholders across finance and treasury teams.

Master of Engineering (M.Tech) - Electronics & Communication Engineering - Indian Institute of Science (IISc) (2018-08 - 2020-06)

Bachelor of Engineering (B.Tech) - Electronics & Communication Engineering - MAKAUT (2014-08 - 2018-06)