Full-Stack ML Engineer with Next.js

Krishi Prashant Shah

+1-530-***-**** **************@*****.*** Davis, CA 95616

LinkedIn GitHub Portfolio

Education

University of California, Davis Sept. 2025 - Present Master of Science in Computer Science, GPA: 3.8/4 Davis, CA SVKM’s Dwarkadas J. Sanghvi College of Engineering Jun. 2021 - May 2025 Bachelor of Engineering in Computer Engineering, CGPA: 9.31/10 Mumbai, India Professional Experience

Student Trainee Apr. 2023 - May 2024

JPMorgan Chase & Co. Mumbai, India

• Architected and shipped 5 full-stack web applications using React.js and Python, collaborating across 7-person cross-functional teams in an Agile environment to deliver production-ready solutions for 3 internal stakeholders/business units.

• Designed and deployed prediction models (risk scoring, fraud detection) integrated into 6 automation workflows, reducing manual processing time by 15% across daily transactions/records. Junior Next JS Developer Aug. 2023 - Jun. 2024

KenmarkItan Solutions Mumbai, India

• Built and deployed 3 production travel partner websites using Next.js, React, Node.js, and MongoDB, architecting CI/CD pipelines that cut deployment time by 40% across 10+ releases.

• Engineered a Next.js music player integrating 6 backend APIs for dynamic streaming, expanding feature coverage by 30% and supporting 5k+ monthly active users.

Projects

Lily Twilio, Deepgram, Claude, ElevenLabs, PostgreSQL, FastAPI Apr. 2026

• Engineered a voice-first maternal triage agent over standard phone lines (PSTN), integrating sub-second streaming STT/TTS pipelines with a deterministic ACOG-grounded rules engine to eliminate LLM hallucination in clinical escalation paths.

• Architected a three-tier triage system (HANDLE, HAND-UP, HAND-OFF) with real-time Twilio conferencing to physicians and 911, backed by a Pinecone vector memory layer for longitudinal patient context across calls. Advanced Machine Learning for Rice Quality Phenotyping Python, Scikit-Learn, Ensemble Learning Jan. 2026

• Engineered a heterogeneous stacked generalization framework (GBR, SVR, Ridge) to fuse 70+ multi-modal physicochemical vectors, including rheological RVA and mechanical TPA data, to decode latent consumer preferences across 1,170+ sensory data points.

• Orchestrated a rigorous Leave-One-Group-Out (LOGO) validation pipeline and GridSearchCV to optimize a 24-fold hyperparameter hyperspace, achieving variety-agnostic predictive stability across 30+ unseen rice cultivars.

Prometheus 2 Next.js, OpenAI, LangChain Sept. 2025

• Engineered a Multi-Modal AI Platform for text, code, music, and image generation serving 100+ active users; reduced response latency by 25%.

• Implemented role-based access control and RESTful API architecture ensuring scalability and security for enterprise adoption.

Technical Skills

Languages: C++, Python, JavaScript, SQL, HTML, CSS Frameworks: Next.js, React, Node.js, Flask, MongoDB, PostgreSQL AI/ML: PyTorch, TensorFlow, Scikit-Learn, LangChain, Anthropic Claude API, Pinecone, ChromaDB, Ollama, RAG, Ensemble Learning

Developer Tools: Docker, Git, Postman, VS Code, Twilio, Deepgram Research & Community

Paper: Published ”Evaluating News Authenticity Using the Hits Algorithm” in Educational Administration: Theory and Practice; improved content accuracy by 25% using authority/hub scores. Paper: ”Lightweight CNN Model for Cotton Leaf Disease Classification” - Achieved 96% accuracy for smart agriculture on edge devices.

Open Source: Contributed to OSSPREY, Apache Resilient DB and many other open source programs Leadership: Google Cloud Arcade Facilitator (guided 100+ students); Placement Coordinator at DJSCE.

Contact this candidate