Jan 2025 – May 2025 · 4 mos · West Lafayette, Indiana, United States · On-site

• Developed patented spectroscopic imaging (MRSI) signal-processing algorithms that cut scan times by ~30%, enabling earlier cancer treatment response detection.
• Applied novel phase-incrementing MRSI methods in Python/MATLAB, improving biomarker quantification accuracy by 20% in pre-clinical studies.
• Collaborated with engineers on RF coil development and multimodal imaging workflows, enhancing tumor immune-cell tracking for translational cancer research.

Jan 2025 – May 2025 · 4 mos · West Lafayette, Indiana, United States · On-site

• Processed and normalized large-scale spatial transcriptomics datasets (CosMx SMI, 1.5GB+) in R/Seurat, enabling downstream cell-type annotation and tumor microenvironment modeling.
• Built CNN models on annotated Seurat data to predict tumor marker expression from cell neighborhoods, achieving 85% accuracy in early validation.
• Engineered k-nearest neighbor pipelines in R to quantify tumor–immune cell interactions, reducing feature extraction time by 40% compared to manual methods.
• Collaborated with researchers at MD Anderson Cancer Center to interpret results, guiding experimental design for ongoing cancer immunology studies.

Aug 2024 – Jan 2025 · 5 mos · West Lafayette, Indiana, United States · On-site

• Designed Python pipelines in Amazon Kinesis to synchronize 4+ sensor streams at a common sampling rate, reducing
• analysis time by 20 minutes per run across 8+ tests.
• Enhanced sensor correlation to over 90% by constructing regression models and visualizations in Scikit-learn and Matplotlib,
• driving product design decisions and securing Q3 funding.
• Validated sensor performance with hypothesis testing and causal inference in R, ensuring reliability of results
• Guided Capstone senior project teams in applying machine learning techniques to model product usage patterns and helped a class of 200 with their client presentations.

Jun 2024 – Aug 2024 · 2 mos · Fort Wayne, Indiana, United States · Hybrid

• Achieved 85% vendor cost reduction by developing predictive models on 2M+ transactions in PySpark to segment customers
• with 94% tagging accuracy.
• Automated ETL workflows in SQL and PySpark on 6 months of transaction data, cutting data load times by 2+ hours.
• Recommended $50K annual cloud savings by evaluating infrastructure and presenting findings to leadership

Jan 2024 – Apr 2024 · 3 mos · West Lafayette, Indiana, United States

• Led supervised fine-tuning of a custom 7B LLM in PyTorch/Transformers with LoRA adapters on 10K+ chemical safety
• documents, cutting training costs by 35% and achieving GPT-3.5–level task accuracy.
• Developed an LLM-powered Q&A pipeline with FAISS that automated shipment label creation and document review,
• saving 100+ worker hours per week and improving compliance efficiency.
• Scaled shipment label automation from a single-vendor rule-based system to LLM-driven workflows supporting 5+ vendors,
• eliminating vendor lock-in and enabling adoption across 4 global safety teams.

Oct 2023 – Aug 2024 · 10 mos · West Lafayette, Indiana, United States · On-site

• Improved annotation throughput by 40% by integrating LLMs and GenAI tools for tissue-specific labeling across MSI datasets of 1.5GB+ each.
• Increased detection accuracy of molecular patterns by 30% through machine learning and computer vision applied to multimodal brain imaging data.
• Built 3D molecular maps combining MSI and optical signals, accelerating biomarker discovery for the NIH BRAIN Initiative, and presented results at ASMS 2024.

Jun 2022 – Aug 2022 · 2 mos · Pune, Maharashtra, India

• Implemented unsupervised machine learning techniques for time series anomaly detection on 2TB of unstructured data in
• Python, reducing total process time by 20%.
• Processed 10TB of trading data in Hadoop using Spark to remodel and visualize 16 previously inaccessible datasets, enabling
• 500+ end clients to track trade impact on liquidity.
• Delivered results to the team’s global head and authored an executive summary detailing value proposition and strategy for senior leadership and clients

Jan 2022 – Feb 2022 · 1 mo · Bhopal, Madhya Pradesh, India

• Built a machine learning pipeline for network intrusion detection on labeled flow data, improving malicious traffic detection accuracy by 35%.
• Engineered structured datasets from raw traffic logs by extracting features such as packet size, duration, and flow type, then trained classification models (Random Forest, SVM) to benchmark detection rates.
• Identified enterprise vulnerabilities by visualizing attack patterns with confusion matrices and time-series plots, supporting integration of the pipeline into a simulated corporate network.

May 2021 – Jun 2022 · 1 yr 1 mo · India · Remote

• Authored deep learning tutorials and sample projects that reached thousands of readers, helping students grasp practical model-building concepts.
• Created beginner-friendly articles on topics such as transfer learning and model deployment, making advanced methods accessible to learners.
• Collaborated with the editorial team to review and refine content, improving clarity, accuracy, and reproducibility for educational resources.