Engineer Intern Generative Ai

Amrutha Kanakatte Ravishankar

201-***-**** *******************@*******.*** linkedin.com/in/amrutha-kanakatte-ravishankar-/ github.com/harithsya24 harithsya24.github.io/profile/ Technical Skills

Programming: Python, C++, SQL, JavaScript

ML/AI Frameworks: PyTorch, TensorFlow, Transformers, RAG Systems, Computer Vision, NLP, Generative AI Quantitative & Statistical Skills: Probability Theory, Statistical Analysis, Time Series, Hypothesis Testing, A/B Testing Cloud & DevOps: AWS (SageMaker, Lambda, EC2, CodePipeline), GCP (Vertex AI), Docker, Kubernetes, Terraform Data & Tools: MySQL, MongoDB, Elasticsearch, FAISS, Redis, FastAPI, Git, NumPy, Pandas, Matplotlib Experience

PONS June 2025 – Present

Artificial Intelligence Research Engineer Intern Newark, NJ

• Engineered solution for critical shortage of diverse training data for breast cancer CAD systems by architecting and deploying a SPADE pix2pix GAN pipeline (3-person team, 4 months), generating anatomically precise ultrasound scans that expanded dataset diversity by 40% and enabled more robust model generalization across patient demographics.

• Eliminated quality bottlenecks in synthetic medical imaging by optimizing conditional GAN architecture with feature-preserving conditioning mechanisms, achieving 18% FID reduction, 12% SSIM increase, and 9% PSNR boost, directly enabling production deployment of CAD models with improved diagnostic accuracy. Stevens School of Business May 2025 – Present

Graduate Research Assistant - NLP Hoboken, NJ

• Solved lack of standardized evaluation metrics for healthcare LLM applications by developing and releasing VECB benchmark (12,000+ question-query pairs across 7 query classes), establishing industry-standard evaluation framework now adopted by 3 research institutions for text-to-NoSQL generation assessment.

• Tackled LLM hallucination challenges in clinical settings by designing and implementing hierarchical RAG framework with medical knowledge base integration, achieving 17% accuracy improvement and 22% hallucination reduction over GPT-4o, enabling reliable clinical decision support for healthcare practitioners. FutureWave Aug. 2021 – May 2022

Software Engineer Intern Mysuru, India

• Resolved platform stability issues affecting 50K+ monthly active users by spearheading development of comprehensive automated testing suites and CI/CD pipelines (6-person engineering team), achieving 20% improvement in platform reliability and eliminating critical service interruptions.

• Accelerated product delivery timelines by refactoring cross-platform codebase and optimizing backend algorithms, reducing technical debt by 35% and improving runtime performance by 15% for global B2B customer base spanning 12 countries. Publications & Research

Comparative Analysis of Transformer Architectures for Intrusion Detection Systems Aug 2024 IEEE NMITCON 2024 - Led research demonstrating 23% improvement in attack detection accuracy Deep Learning Approaches for Audio Source Separation in Karaoke Applications Jun 2023 IJIRAE, Vol. 10, Issue 06 - First author

Education

Stevens Institute of Technology Jan. 2024 – Dec. 2025 Master of Science in Computer Science, GPA: 3.86/4.0 Hoboken, NJ Relevant Coursework: Machine Learning, Statistical Machine Learning, Deep Learning, Natural Language Processing, Generative AI, Data Mining

JSS Science and Technology University Aug. 2019 – Aug. 2023 Bachelor of Engineering in Electronics and Communication, GPA: 3.58/4.0 Mysuru, India Projects

CloudMart: Multi-Cloud E-commerce Platform AWS, GCP, Azure, Terraform, GitHub January 2025

• Architected fault-tolerant multi-cloud e-commerce infrastructure (AWS/GCP/Azure) with automated CI/CD pipelines, achieving 99.9% uptime during 3x peak traffic and reducing deployment time by 79% (2 hours 25 minutes). Agentic Medical RAG System for Clinical Decision Support RAG, FAISS, PubMed API December 2024

• Built FAISS-indexed clinical decision support system processing 1.2M medical articles with sub-300ms response times and 95% precision, reducing physician research time by 87% (15min 2min per query). Certifications & Recognition

AWS Certified AI Practitioner July 2025

Quack Hack Hackathon Winner Innovation Recognition

Contact this candidate