Machine Learning Engineer - GenAI & Computer Vision

DHRUVIL GORASIYA

New York, NY +1-623-***-**** *******.*@*********.*** LinkedIn GitHub Portfolio SUMMARY

Machine Learning Engineer with 4+ years of experience designing, training, and deploying production ML systems across payments fraud, healthcare NLP, Product and retail computer vision. Deep expertise in deep learning, NLP, computer vision, and generative / agentic AI (LLM fine-tuning, RAG, multi-agent orchestration). Proven record of shipping low-latency models at scale and converting ambiguous problems into measurable revenue and risk outcomes. SKILLS

Languages: Python, SQL, Scala, Java, C++, R, Bash

ML & Deep Learning: PyTorch, TensorFlow, Keras, Scikit-learn, XGBoost, LightGBM, CNNs, RNNs / LSTMs, Transformers, Bayesian methods

Generative & Agentic AI: LLM fine-tuning (LoRA / PEFT), RAG, LangChain, LangGraph, vector databases (FAISS, Pinecone), prompt engineering, multi-agent orchestration, MCP, Hugging Face NLP & Computer Vision: NER, text classification, summarization, embeddings, BERT, spaCy; object detection (YOLO), segmentation, OCR / Document AI, Vision Transformers, OpenCV MLOps & Cloud: AWS (SageMaker), GCP (Vertex AI, BigQuery), Docker, Kubernetes, MLflow, Kubeflow, Airflow, CI / CD, model monitoring

Data & Experimentation: Spark, Kafka, Snowflake, Pandas, feature stores, ETL, A / B testing, statistical inference, drift detection WORK EXPERIENCE

Stripe

Machine Learning Engineer, Contract Payments, Fraud & Risk Jun 2025 – Present Machine Learning Engineer, Intern Jan 2025 – May 2025

• Engineered and deployed gradient-boosted and deep-learning fraud-detection features into Radar’s real-time scoring pipeline, informing risk decisions on millions of daily transactions at sub-100 ms latency.

• Built 40+ behavioral and network features feeding Stripe’s transformer-based payments foundation model, lifting fraud recall by 2.4% on a high-risk merchant segment while holding false-positive rate flat.

• Designed an LLM-powered agent that converts plain-English fraud policies into validated, executable rule code, cutting analyst rule-authoring time from 30 minutes to under 3 minutes across 200+ rules.

• Optimized authorization-rate models powering Adaptive Acceptance, recovering an additional 1.1% of legitimate payments across a high-volume card segment without raising fraud exposure.

• Shipped a chargeback-prediction model into Smart Disputes that automates evidence selection, improving dispute win-rate by 9% and reducing manual analyst review volume.

• Established offline and online evaluation, drift monitoring, and A / B testing with cross-functional risk and data-engineering partners, and presented business impact to senior staff under PCI-DSS and SCA / 3DS compliance. Vivma Software Inc.

Machine Learning Engineer Healthcare & Retail AI on Google Cloud Oct 2021 – Jul 2023

• Productionized a clinical NLP entity-extraction pipeline on Vertex AI and Healthcare NL API, raising extraction F1 from 0.74 to 0.88 and reducing manual chart-review effort by 40% for a US healthcare client.

• Developed computer-vision OCR and Document-AI models to digitize unstructured medical and invoice documents, processing 1M+ pages monthly at 96% field-level accuracy.

• Trained deep-learning demand-forecasting models for a retail / CPG client, lowering forecast error (MAPE from 18% to 11%) and improving inventory allocation across 500+ SKUs.

• Owned end-to-end MLOps: containerized training and serving with Docker and Kubeflow / Vertex Pipelines, added drift detection and automated retraining, cutting model-refresh time by 60%.

• Fine-tuned an early LLM-based clinical-summarization prototype on Vertex AI, reducing clinician documentation time by 35% in a pilot across 3 hospital units.

• Led client-facing stakeholder workshops, translated requirements into solution designs, and enforced HIPAA and Responsible- AI bias checks across deployments.

Vivma Software Inc.

Data Science Intern Jan 2021 – Sep 2021

• Prototyped CNN and transformer baselines for image- and text-classification proofs of concept, accelerating client demos and informing two production roadmaps.

• Benchmarked 6 pretrained CV and NLP architectures, mapping accuracy-latency trade-offs that guided model-selection decisions across two client engagements.

• Created interactive evaluation dashboards in Python and Looker Studio, giving stakeholders clear visibility into model performance and data-quality trends.

• Automated data-ingestion and feature-engineering pipelines in Python and BigQuery, reducing dataset-prep time by 50% and earning promotion to Machine Learning Engineer.

EDUCATION

Arizona State University Tempe, AZ

Master of Science, Computer Science Aug 2023 – May 2025 KEY PROJECTS

• Multi-Agent RAG Assistant: LangGraph agents over a FAISS store with tool-calling and MCP, hitting 92% answer-grounding on 10K docs.

• Real-Time Vision Defect Detection: YOLO / ViT model for manufacturing defects, achieving 0.91 mAP at 25 FPS inference. CERTIFICATIONS

Google Cloud Professional Machine Learning Engineer AWS Certified Machine Learning – Specialty ACHIEVEMENTS

• Recognized for impactful contributions to the field of plasma research by presenting groundbreaking work at the prestigious International Conference on Numerical Simulation of Plasmas (ICNSP), held in Japan in 2022.

Contact this candidate