Hi, I'm Sunny!
Hi, I'm Sunny!
AI & ML Specialist
Cloud-Native Innovator
Full Stack Developer
DevOps Automation Pro
AI & ML Specialist
Cloud-Native Innovator
Full Stack Developer
DevOps Automation Pro
About Me
Passionate about technology, I believe the future is shaped by those who build it. I specialize in crafting innovative solutions, blending AI, machine learning, and full-stack development. Code isn't just a tool for me—it's a language that speaks to the future and helps solve real-world challenges.
Passionate about technology, I believe the future is shaped by those who build it. I specialize in crafting innovative solutions, blending AI, machine learning, and full-stack development. Code isn't just a tool for me—it's a language that speaks to the future and helps solve real-world challenges.
Over the years, I’ve embraced this language to shape projects in AI, NLP, cloud computing, and web development. From improving business processes through machine learning models to building scalable systems, my journey is a fusion of curiosity, innovation, and a drive to push beyond boundaries.
Over the years, I’ve embraced this language to shape projects in AI, NLP, cloud computing, and web development. From improving business processes through machine learning models to building scalable systems, my journey is a fusion of curiosity, innovation, and a drive to push beyond boundaries.
My work goes beyond creation—it's about envisioning what’s next. As technology evolves, so do the challenges. Whether it’s developing intelligent systems or designing future-ready architectures, I’m constantly driven by one question: How can I build something that stands the test of time?
My work goes beyond creation—it's about envisioning what’s next. As technology evolves, so do the challenges. Whether it’s developing intelligent systems or designing future-ready architectures, I’m constantly driven by one question: How can I build something that stands the test of time?
I Believe,
Curiosity leads to creation,
Creation drives change.
Passionate about technology, I believe the future is shaped by those who build it. I specialize in crafting innovative solutions, blending AI, machine learning, and full-stack development. Code isn't just a tool for me—it's a language that speaks to the future and helps solve real-world challenges.
Passionate about technology, I believe the future is shaped by those who build it. I specialize in crafting innovative solutions, blending AI, machine learning, and full-stack development. Code isn't just a tool for me—it's a language that speaks to the future and helps solve real-world challenges.
Over the years, I’ve embraced this language to shape projects in AI, NLP, cloud computing, and web development. From improving business processes through machine learning models to building scalable systems, my journey is a fusion of curiosity, innovation, and a drive to push beyond boundaries.
Over the years, I’ve embraced this language to shape projects in AI, NLP, cloud computing, and web development. From improving business processes through machine learning models to building scalable systems, my journey is a fusion of curiosity, innovation, and a drive to push beyond boundaries.
My work goes beyond creation—it's about envisioning what’s next. As technology evolves, so do the challenges. Whether it’s developing intelligent systems or designing future-ready architectures, I’m constantly driven by one question: How can I build something that stands the test of time?
My work goes beyond creation—it's about envisioning what’s next. As technology evolves, so do the challenges. Whether it’s developing intelligent systems or designing future-ready architectures, I’m constantly driven by one question: How can I build something that stands the test of time?
Tracing the Path of Enovation
Tracing the Path of Enovation
Tracing the Path of Enovation
MY WORK EXPERIENCE
MY WORK EXPERIENCE
MY WORK EXPERIENCE
AI Engineer
FortyAU
MAY 2025—Present | Nashville, Tennessee, U.S.
MAY 2025—Present
Nashville, Tennessee, U.S.
+ Built LLM-powered Text-to-SQL/Lucene pipelines for enterprise and mid-market clients, handling follow-ups using the last 5 interactions and conversation summaries, supporting 10k+ queries/month. Developed LangGraph agent workflows with nodes for guardrails, clarification, intent classification, query rewriting, RAG retrieval, and answer generation, incorporating a micro-fine-tuned GPT-4o model for complex queries with automatic routing.
+ Optimized performance via multi-level caching: query-level cache delivered near-instant responses (~0.01 s) for repeated/semantic queries, and step-level LRU memoization cut first-time latency ~72% (45 s → 12.5 s). Reduced token usage ~89% (4,500 → 500 tokens) using intelligent table selection and lightweight planning with GPT-4o-mini.
+ Built Streamlit dashboards (OpenAI/DeepEval/RAGAS) tracking answer relevance, context accuracy, and retrieval quality, integrated LangSmith and Arize Phoenix for observability, improving QA efficiency ~30%. Leveraged AI tools (Cursor, Cline, Claude) to automate code, streamline agent workflows, and boost development efficiency ~33%, enabling rapid prototyping and scalable deployment.
Generative AI Engineer Intern
Generative AI Engineer Intern
Austrian Post
Austrian Post
AUG 2024—DEC 2024 | Remote, U.S.A
AUG 2024—DEC 2024
Remote, U.S.A
+ Fine-tuned GPT-4o on time-series logs using LoRA adapters, increasing inference throughput by 42% while maintaining low compute overhead, enabling faster and more efficient data-driven decision-making.
+ Developed a responsive Vue.js interface backed by FastAPI, integrating AWS Lambda triggers and DynamoDB TTLs, reducing response latency to 57ms and enhancing real-time system interaction.
+ Deployed CrewAI-based agents to manage complex multi-step reasoning tasks, orchestrating workflows with Docker, Kubernetes Jobs, and Snowflake alerts across two cloud-monitored regional clusters for scalable and reliable operations.
Research and Development Intern
Vanderbilt University
Vanderbilt University
AUG 2024—DEC 2024 | Nashville, Tennessee, U.S.
AUG 2024—DEC 2024
Nashville, Tennessee, U.S.
Supervisor: Dr. Daniel Balasubramanian
Intelligent LLM-Powered Fault Detection for DARPA BPL Workflows
+ Developed an LLM-powered fault detection system using PyTorch MLPs and SQL anomaly detection on 14,200 BPMN events, autonomously identifying and preventing critical workflow faults for DARPA’s BPL program.
+ Engineered a RAG system with FAISS cosine indexing, PostgreSQL full-text search, and LangGraph agents, reducing NLP latency by ~20% and applying advanced NLP techniques (dependency parsing, NER, coreference resolution) for precise entity mapping across BPMN diagrams, XML, and PDFs.
+ Automated training and evaluation pipelines with Airflow DAGs, Pandas feature extractors, and MLflow logging, while collaborating with Leidos to meet DoD requirements, ensuring scalable, reproducible workflows and mission-critical fault prevention.
Website Developer
Website Developer
7 Code Solutions
7 Code Solutions
SEP 2021—FEB 2023 | Gandhinagar, Gujarat, India
SEP 2021—FEB 2023
Gandhinagar, Gujarat, India
Personalization Engine Enhancement
+ Built a Python-BERT recommendation engine using time-series lag features and attention pooling, reducing user churn by 25%. Implemented Sentence-BERT for NLP tasks, achieving 15% better content tagging accuracy and 90% precision in sentiment analysis.
+ Integrated GraphQL APIs with Node.js and Express, and orchestrated pipelines with Kubernetes, ensuring scalable and maintainable backend services.
+ Developed a modular React frontend with HTML5, CSS3, and JavaScript (ES6+) for responsive and maintainable user interfaces.
AI-driven A/B Testing Platform
+ Designed and implemented an A/B testing platform using React for an interactive dashboard and Redux for state management, improving conversion funnel efficiency by 15%.
+ Applied multi-armed bandit algorithms (epsilon-greedy, Thompson sampling) to optimize traffic distribution, supporting data-driven decision-making.
+ Conducted statistical evaluation using SQL UDFs, PostgreSQL analytics, and Welch’s t-test, ensuring robust measurement of experiment outcomes.
AI Engineer
MAY 2025—Present
Nashville, Tennessee, U.S.
FortyAU
+ Built LLM-powered Text-to-SQL/Lucene pipelines for enterprise and mid-market clients, handling follow-ups using the last 5 interactions and conversation summaries, supporting 10k+ queries/month. Developed LangGraph agent workflows with nodes for guardrails, clarification, intent classification, query rewriting, RAG retrieval, and answer generation, incorporating a micro-fine-tuned GPT-4o model for complex queries with automatic routing.
+ Optimized performance via multi-level caching: query-level cache delivered near-instant responses (~0.01 s) for repeated/semantic queries, and step-level LRU memoization cut first-time latency ~72% (45 s → 12.5 s). Reduced token usage ~89% (4,500 → 500 tokens) using intelligent table selection and lightweight planning with GPT-4o-mini.
+ Built Streamlit dashboards (OpenAI/DeepEval/RAGAS) tracking answer relevance, context accuracy, and retrieval quality, integrated LangSmith and Arize Phoenix for observability, improving QA efficiency ~30%. Leveraged AI tools (Cursor, Cline, Claude) to automate code, streamline agent workflows, and boost development efficiency ~33%, enabling rapid prototyping and scalable deployment.
Research and Development Intern
AUG 2024—DEC 2024
Nashville, Tennessee, U.S.
Vanderbilt University
Supervisor: Dr. Daniel Balasubramanian
Intelligent LLM-Powered Fault Detection for DARPA BPL Workflows
+ Developed an LLM-powered fault detection system using PyTorch MLPs and SQL anomaly detection on 14,200 BPMN events, autonomously identifying and preventing critical workflow faults for DARPA’s BPL program.
+ Engineered a RAG system with FAISS cosine indexing, PostgreSQL full-text search, and LangGraph agents, reducing NLP latency by ~20% and applying advanced NLP techniques (dependency parsing, NER, coreference resolution) for precise entity mapping across BPMN diagrams, XML, and PDFs.
+ Automated training and evaluation pipelines with Airflow DAGs, Pandas feature extractors, and MLflow logging, while collaborating with Leidos to meet DoD requirements, ensuring scalable, reproducible workflows and mission-critical fault prevention.
Generative AI Engineer Intern
AUG 2024—DEC 2024
Remote, U.S.
Austrian Post
+ Fine-tuned GPT-4o on time-series logs using LoRA adapters, increasing inference throughput by 42% while maintaining low compute overhead, enabling faster and more efficient data-driven decision-making.
+ Developed a responsive Vue.js interface backed by FastAPI, integrating AWS Lambda triggers and DynamoDB TTLs, reducing response latency to 57ms and enhancing real-time system interaction.
+ Deployed CrewAI-based agents to manage complex multi-step reasoning tasks, orchestrating workflows with Docker, Kubernetes Jobs, and Snowflake alerts across two cloud-monitored regional clusters for scalable and reliable operations.
Website Developer
SEP 2021—FEB 2023
Gandhinagar, Gujarat, India
7 Code Solutions
Personalization Engine Enhancement
+ Built a Python-BERT recommendation engine using time-series lag features and attention pooling, reducing user churn by 25%. Implemented Sentence-BERT for NLP tasks, achieving 15% better content tagging accuracy and 90% precision in sentiment analysis.
+ Integrated GraphQL APIs with Node.js and Express, and orchestrated pipelines with Kubernetes, ensuring scalable and maintainable backend services.
+ Developed a modular React frontend with HTML5, CSS3, and JavaScript (ES6+) for responsive and maintainable user interfaces.
AI-driven A/B Testing Platform
+ Designed and implemented an A/B testing platform using React for an interactive dashboard and Redux for state management, improving conversion funnel efficiency by 15%.
+ Applied multi-armed bandit algorithms (epsilon-greedy, Thompson sampling) to optimize traffic distribution, supporting data-driven decision-making.
+ Conducted statistical evaluation using SQL UDFs, PostgreSQL analytics, and Welch’s t-test, ensuring robust measurement of experiment outcomes.
Tech that fuels the future
MY STACK
The backbone of my tech stack, Python empowers me to create efficient automation scripts and advanced machine learning models, ensuring reliability and scalability in every project.
Python
The backbone of my tech stack, Python empowers me to create efficient automation scripts and advanced machine learning models, ensuring reliability and scalability in every project.
Python
Tap to flip
The backbone of my tech stack, Python empowers me to create efficient automation scripts and advanced machine learning models, ensuring reliability and scalability in every project.
Python
Tap to flip
The backbone of my tech stack, it empowers me to create efficient automation scripts and advanced machine learning models, ensuring reliability and scalability in every project.
TensorFlow
The backbone of my tech stack, it empowers me to create efficient automation scripts and advanced machine learning models, ensuring reliability and scalability in every project.
TensorFlow
Tap to flip
The backbone of my tech stack, it empowers me to create efficient automation scripts and advanced machine learning models, ensuring reliability and scalability in every project.
TensorFlow
Tap to flip
I enjoy its flexibility and intuitive design in developing dynamic AI solutions. Its ease of use allows me to iterate quickly and achieve optimal performance.
PyTorch
I enjoy its flexibility and intuitive design in developing dynamic AI solutions. Its ease of use allows me to iterate quickly and achieve optimal performance.
PyTorch
Tap to flip
I enjoy its flexibility and intuitive design in developing dynamic AI solutions. Its ease of use allows me to iterate quickly and achieve optimal performance.
PyTorch
Tap to flip
I use it to craft interactive user experiences, leveraging its component-based architecture to deliver fast and responsive web applications that engage users seamlessly.
ReactJS
I use it to craft interactive user experiences, leveraging its component-based architecture to deliver fast and responsive web applications that engage users seamlessly.
ReactJS
Tap to flip
I use it to craft interactive user experiences, leveraging its component-based architecture to deliver fast and responsive web applications that engage users seamlessly.
ReactJS
Tap to flip
My go-to for robust cloud architecture. I design scalable solutions, from storage to serverless functions, ensuring high availability and resilience in my applications.
AWS
My go-to for robust cloud architecture. I design scalable solutions, from storage to serverless functions, ensuring high availability and resilience in my applications.
AWS
Tap to flip
My go-to for robust cloud architecture. I design scalable solutions, from storage to serverless functions, ensuring high availability and resilience in my applications.
AWS
Tap to flip
To revolutionize my development workflow by containerizing applications. This ensures consistency across environments and simplifies deployment for any project I undertake.
Docker
To revolutionize my development workflow by containerizing applications. This ensures consistency across environments and simplifies deployment for any project I undertake.
Docker
Tap to flip
To revolutionize my development workflow by containerizing applications. This ensures consistency across environments and simplifies deployment for any project I undertake.
Docker
Tap to flip
Notion helps me keep my projects organized. I use it for project management, task tracking, and as a central hub for documentation, ensuring that everything from design notes to project timelines is in one place.
Kubernetes
Notion helps me keep my projects organized. I use it for project management, task tracking, and as a central hub for documentation, ensuring that everything from design notes to project timelines is in one place.
Kubernetes
Tap to flip
Notion helps me keep my projects organized. I use it for project management, task tracking, and as a central hub for documentation, ensuring that everything from design notes to project timelines is in one place.
Kubernetes
Tap to flip
I use it to implement Infrastructure as Code, enabling precise management of cloud resources while automating deployment processes for maximum efficiency.
Terraform
I use it to implement Infrastructure as Code, enabling precise management of cloud resources while automating deployment processes for maximum efficiency.
Terraform
Tap to flip
I use it to implement Infrastructure as Code, enabling precise management of cloud resources while automating deployment processes for maximum efficiency.
Terraform
Tap to flip
It streamlines my data workflows, enabling me to schedule and manage complex data pipelines effortlessly, ensuring timely and efficient task execution.
Apache Airflow
It streamlines my data workflows, enabling me to schedule and manage complex data pipelines effortlessly, ensuring timely and efficient task execution.
Apache Airflow
Tap to flip
It streamlines my data workflows, enabling me to schedule and manage complex data pipelines effortlessly, ensuring timely and efficient task execution.
Apache Airflow
Tap to flip
I enhance my machine learning lifecycle management, tracking experiments and deployments efficiently with it to ensure reproducibility and collaboration across teams.
MLFlow
I enhance my machine learning lifecycle management, tracking experiments and deployments efficiently with it to ensure reproducibility and collaboration across teams.
MLFlow
Tap to flip
I enhance my machine learning lifecycle management, tracking experiments and deployments efficiently with it to ensure reproducibility and collaboration across teams.
MLFlow
Tap to flip
It empowers me to handle unstructured data at scale, providing the flexibility needed to manage large datasets while ensuring high performance and agility.
MongoDB
It empowers me to handle unstructured data at scale, providing the flexibility needed to manage large datasets while ensuring high performance and agility.
MongoDB
Tap to flip
It empowers me to handle unstructured data at scale, providing the flexibility needed to manage large datasets while ensuring high performance and agility.
MongoDB
Tap to flip
It provides a powerful relational database solution for managing structured data, offering advanced querying capabilities to support data-driven decision-making.
PostgreSQL
It provides a powerful relational database solution for managing structured data, offering advanced querying capabilities to support data-driven decision-making.
PostgreSQL
Tap to flip
It provides a powerful relational database solution for managing structured data, offering advanced querying capabilities to support data-driven decision-making.
PostgreSQL
Tap to flip
Version control through Git is crucial to my workflow. With GitHub, I manage collaborative development, track changes, and ensure code integrity across multiple projects and teams.
Git/GitHub
Version control through Git is crucial to my workflow. With GitHub, I manage collaborative development, track changes, and ensure code integrity across multiple projects and teams.
Git/GitHub
Tap to flip
Version control through Git is crucial to my workflow. With GitHub, I manage collaborative development, track changes, and ensure code integrity across multiple projects and teams.
Git/GitHub
Tap to flip
It automates my continuous integration processes, allowing me to test and deploy code seamlessly, ensuring that every update meets high-quality standards.
Jenkins
It automates my continuous integration processes, allowing me to test and deploy code seamlessly, ensuring that every update meets high-quality standards.
Jenkins
Tap to flip
It automates my continuous integration processes, allowing me to test and deploy code seamlessly, ensuring that every update meets high-quality standards.
Jenkins
Tap to flip
It enhances my CI/CD pipeline, automating testing and deployment workflows to accelerate development cycles and deliver reliable software to production faster.
CircleCI
It enhances my CI/CD pipeline, automating testing and deployment workflows to accelerate development cycles and deliver reliable software to production faster.
CircleCI
Tap to flip
It enhances my CI/CD pipeline, automating testing and deployment workflows to accelerate development cycles and deliver reliable software to production faster.
CircleCI
Tap to flip
Here's How I Solve Challenges
PROJECTS
Turning Data Into Discovery
Research Papers
An ML-Based Approach to Transform Stimuli into Characteristic Oscillatory Activity of the Beta Band
Journal/Conference:
Authors:
Duration:
RIACT 2022
Sunny Bhatt, Aditi Jindal
6 months
An ML-Based Approach to Transform Stimuli into Characteristic Oscillatory Activity of the Beta Band
Journal/Conference:
Authors:
Duration:
RIACT 2022
Sunny Bhatt, Aditi Jindal
6 months
Implementation Of Crop Disease Detection System
Using Flutter
Journal/Conference:
Authors:
Duration:
RIACT 2022
Sunny Bhatt, Shivam Hingu
6 months
Implementation Of Crop Disease Detection System
Using Flutter
Journal/Conference:
Authors:
Duration:
RIACT 2022
Sunny Bhatt, Shivam Hingu
6 months
An ML-Based Approach to Transform Stimuli into Characteristic Oscillatory Activity of the Beta Band
Journal/Conference:
Authors:
Duration:
RIACT 2022
Sunny Bhatt, Aditi Jindal
6 months
Implementation Of Crop Disease Detection System
Using Flutter
Journal/Conference:
Authors:
Duration:
RIACT 2022
Sunny Bhatt, Shivam Hingu
6 months
Curiosities Beyond Code
INTERESTS AND HOBBIES
Bug–Squashing
Data Diving
Designing Intelligent Systems
Bug–Squashing
Data Diving
Designing Intelligent Systems
STARGAZING
HIKING
HIKING
I SING TOO !!
I SING TOO !!
Let's Talk!
