Nirav Madhani
Machine Learning Engineer
About Me
I'm a Software Engineer passionate about building intelligent applications and scalable cloud infrastructure. With a background in AI and full-stack development, I enjoy tackling complex problems and turning ideas into robust, production-ready solutions.
My experience spans from developing RAG-powered chatbots and optimizing Java microservices to designing scalable systems on Azure. I'm always eager to learn new technologies and contribute to innovative projects.
Experience
Application Developer II
Sep '23 - PresentARGO DATA, Richardson, TX
- Optimized various API features, saving 182,520 vCPU-minutes annually for over 500,000 users.
- Contributed to setting up Azure Monitor and Alerts, reducing problem detection time by 15 minutes and solving time by 7 minutes per incident.
- Re-architected the application for VNet implementation, mitigating security risks and accelerating database calls by more than 300%.
- Led the development of an LLM-powered AI chatbot for surveys using the LangChain Agents framework, achieving 3x faster development.
- Performed LoRA based finetuning, yielding a 10% average improvement across all benchmarks.
AzureLangChainPythonLLMOpsMLOpsFinetuningAgentsMonitoringOptimizationJavaAgile
Application Developer I
Sep '22 - Aug '23ARGO DATA, Richardson, TX
- Increased UI performance by 20% and averted security risks worth over $200,000 by upgrading React libraries and migrating to MUI v5.
- Helped increase efficiency by 6x for generating high-quality LLM finetuning data by creating a React-based data visualization dashboard.
- Migrated applications from monolithic architecture to Azure microservices, reducing infrastructure costs by $20,000 annually.
AzurePythonLLMOpsEDAFinetuningAutomationReactPipelinesDockerContainers
Full-Time Research Intern
Jan '22 - May '22Indian Space Research Organization
- Optimized a hyperspectral super-resolution model, improving processing speed by 30% for large-scale satellite remote sensing.
- Improved image resolution from 20m to 5m, impacting 442 million acres of forest and agricultural land.
- Conducted spectral resampling on over 1TB of data, increasing model accuracy by 8%.
Deep LearningComputer VisionHigh Performance ComputingPyTorchMXNet
Recent Projects
No projects to display.
Education
M.S. in Computer Science
Sep '23 - Dec '25University of Texas, Austin
GPA: 3.75
Coursework: Reinforcement Learning, Deep Learning, Robotics Control, Online Learning & Optimization
B.Tech. in Computer Science & Engineering
Jul '18 - Jun '22Nirma University, India
Coursework: Cloud Computing, Data Mining, Machine Learning, Scientific Computing, Big Data Analytics
Skills
Programming Languages
PythonJavaC++C#VB.NetJavaScript
Databases
MySQLSQL ServerPostgres SQLMongoDBCassandraPinecone
Frameworks
DjangoFlaskSpring BootBootstrapReactNode.js
Cloud Platforms
AWSAzureGCPDockerKubernetesOpenShiftDigital OceanServerless Functions
Machine Learning
LLMsNLPVision-Language ModelsRAGTime-SeriesPyTorchTensorFlowLangChainONNXscikit-learnllama-indexCausalityModel EvaluationMLOpsLLMOps
Misc
GitHubJenkinsGraphQLRESTAgileJiraROSJSON-RPCAgentsChain of ThoughtAPI CompositionBusiness & IT Automation