Headline

Applied Machine Learning Engineer | LLM & Intelligent Systems | Interested in Scientific ML and Multimodal Modelling

About

I am a Machine Learning Engineer focused on building practical AI systems while exploring the application of machine learning in scientific and complex quantitative domains. My background is in applied AI engineering, where I have worked on LLM-based systems, retrieval-augmented pipelines, and multimodal applications. I enjoy designing end-to-end ML systems from experimentation and evaluation to scalable production deployment. Before transitioning fully into AI, I studied chemical engineering, which sparked my interest in modelling real-world physical and data-driven systems. This has gradually led me toward Scientific Machine Learning, Physics-Informed modelling, and learning from multidimensional data. My technical interests currently include: • Scientific and physics-informed ML • LLM and intelligent agent systems • Multimodal representation learning • Computer vision for scientific and satellite data • Scalable ML infrastructure and evaluation I am deeply interested in applications across climate modelling, material science, simulation systems, sports analytics, and digital twins. I am open to applied ML and research-oriented collaborations that involve challenging modelling or intelligent system design problems.

Location

Ahmedabad

Work experience

Artificial Intelligence Consultant at Helix Logic, Jan 2025 - Present · 1 yr 4 mos, San Francisco Bay Area, Compliance Copilot — City of Stanwood • Led end-to-end prototyping and delivery of a Compliance Copilot for the City of Stanwood, US, integrating LLMs with municipal regulatory data to automate compliance review workflows. • Owned project scoping, architecture decisions, and stakeholder communication from brief to delivery. Mental Health Companion with Behavioral Intelligence • Designed and delivered a mental health companion app with behavioral intelligence integration, leveraging LLMs and longitudinal behavioral pattern tracking for personalized intervention support. • Established the AI architecture and led model selection, fine-tuning, and safety evaluation.

Founding Engineer(ML) at Strative, Feb 2025 - Nov 2025 · 10 mos, San Francisco, CA, As Founding Engineer with efficiently small team, this role strengthened my expertise in production-scale AI system design, evaluation-driven model development and bridging research to production gap. • Led end-to-end development of an agentic AI system for auto-liability insurance, automating ~50% of underwriting workflows through autonomous document reasoning and structured data extraction • Designed and deployed scalable production infrastructure using GCP, Kubernetes, FastAPI, Elasticsearch, and MongoDB, enabling reliable and observable AI services • Built evaluation datasets from real customer workflows to continuously validate model accuracy against real-world insurance cases • Enhanced reasoning and retrieval workflows using DSPy and TextGrad, achieving ~97% accuracy compared to human expert baselines • Contributed to technical whitepapers documenting system architecture, component interactions, and measurable business impact

Machine Learning Engineer at Advic Tech, Apr 2024 - Feb 2025 · 11 mos, Hybrid, I worked on building AI-powered document intelligence solutions using retrieval-augmented generation (RAG). I developed a document chat application that enabled interaction with PDFs, DOCX, PPT files, and web URLs, integrating multilingual vector search and text generation using open-source LLMs. I optimized file processing by designing a Redis-based queue, which improved processing speed by 40%. Additionally, I built a scalable backend system using Docker, SQL, and MongoDB, ensuring efficient real-time information retrieval across 1000+ files. My work focused on enhancing AI-driven document processing and information retrieval.

Machine Learning Engineer at Codiste, Nov 2023 - Apr 2024 · 6 mos, Ahmedabad, Gujarat, India, -Stable Diffusion models •Build a pipeline of stable diffusion models using open sourced checkpoints and Lora adapters •Deployed serverless pipeline for text2image generation for integration in apps -RAG based application •Integrated Mongo vector db for storing text embedding for efficient RAG pipline •Integrated a database (Milvus vector db) for storing text embedding for efficient RAG pipline •Conducted iterative experiments for query search with search metric and indexing algorithms like "L2", "cosine" with "HNSW", "IVF_FLAT" -Talking Head • Led a project creating talking heads from still images using a SadTalker model that processes audio and images. •Conducted in-depth research, extensively reviewing numerous papers and code repositories, to identify a state-of-the-art model. •Conducted iterative experiments to optimize hyperparameters, ensuring the model's performance aligns with project objectives. •Integrated OpenAI's TTS for text-to-audio conversion, connecting it seamlessly with the model for video generation. •Deployed model on serverless serving using runpod services -Integration of open source LLM •Evaluated model outputs, considering the training process and techniques such as SFT, RLHF, and DPO. •Utilized libraries such as Hugging Face to implement and streamline the evaluation of open source language models (LLMs) of different sizes and hyperparameters to fit the use case. •Conducted analysis to generate structured output through LLMs. •Performed benchmarking on LLMs, assessing token counts and inference speed. •Conducted benchmark tests on diverse hardware platforms for comprehensive evaluation. •Contributed to prompt engineering to enhance model performance and responsiveness.

Machine Learning Engineer at IQud Tek LLP, Jul 2023 - Oct 2023 · 4 mos, Gandhinagar, Gujarat, India

Education

Master of Science - MS, Artificial Intelligence and Machine Learning, Gujarat University

Associate's degree, Artificial Intelligence, Aptech Learning

Public links

github: https://github.com/yaga-07