Machine Learning Researcher & Software Engineer
Hi, I'm Rohan Sawai
A machine learning engineer who loves building LLM-powered AI agents and scalable systems. I turn research into real-world impact, from deploying transformer models to crafting robust backends.
When I’m not coding, you’ll find me on the soccer field, at the gym, or grinding LeetCode.
I’m currently looking for research opportunities or assistantships in ML/AI—especially anything involving LLMs, agentic workflows, or applied deep learning. If you want to chat about research, code, working out, or even soccer, let’s connect!
Work Experience
My professional journey in software development
Software Developer
Center for Geospatial Information Technology, Virginia TechBlacksburg, VAFeb. 2025 – Present
- •Part of my work at CGIT was to build AI tools that let police and city officials ask questions like "Show me severe crashes in Fairfax County in the last 2 weeks?” and see the results directly on interactive maps.
- •I built AI agents to turn complex natural language queries into optimized multilevel SQL/PostGIS queries, then visualized the answers as map points, heatmaps, and spatial layers using GeoDjango and Leaflet.
- •This work helped transform raw map data into actionable, location-based insights for public safety.
Software Developer
Center for Geospatial Information Technology, Virginia TechBlacksburg, VAFeb. 2025 – Present
- •Part of my work at CGIT was to build AI tools that let police and city officials ask questions like "Show me severe crashes in Fairfax County in the last 2 weeks?” and see the results directly on interactive maps.
- •I built AI agents to turn complex natural language queries into optimized multilevel SQL/PostGIS queries, then visualized the answers as map points, heatmaps, and spatial layers using GeoDjango and Leaflet.
- •This work helped transform raw map data into actionable, location-based insights for public safety.
Software Developer
Center for Geospatial Information Technology, Virginia TechBlacksburg, VAFeb. 2025 – Present
- •Part of my work at CGIT was to build AI tools that let police and city officials ask questions like "Show me severe crashes in Fairfax County in the last 2 weeks?” and see the results directly on interactive maps.
- •I built AI agents to turn complex natural language queries into optimized multilevel SQL/PostGIS queries, then visualized the answers as map points, heatmaps, and spatial layers using GeoDjango and Leaflet.
- •This work helped transform raw map data into actionable, location-based insights for public safety.
MS in Computer Engineering
Virginia TechBlacksburg, VAJan. 2025 – Present
- •GPA: 3.90/4.00
- •While at Citi bank, I realized I wanted to go deeper into ML and LLMs, so I headed to Virginia Tech for my master’s.
- •Now I’m focused on research in LLMs especially agentic workflows, vector DBs and memory component.
- •Coursework: Advanced Machine Learning, Applications of Machine Learning, Compiler Optimizations
MS in Computer Engineering
Virginia TechBlacksburg, VAJan. 2025 – Present
- •GPA: 3.90/4.00
- •While at Citi bank, I realized I wanted to go deeper into ML and LLMs, so I headed to Virginia Tech for my master’s.
- •Now I’m focused on research in LLMs especially agentic workflows, vector DBs and memory component.
- •Coursework: Advanced Machine Learning, Applications of Machine Learning, Compiler Optimizations
MS in Computer Engineering
Virginia TechBlacksburg, VAJan. 2025 – Present
- •GPA: 3.90/4.00
- •While at Citi bank, I realized I wanted to go deeper into ML and LLMs, so I headed to Virginia Tech for my master’s.
- •Now I’m focused on research in LLMs especially agentic workflows, vector DBs and memory component.
- •Coursework: Advanced Machine Learning, Applications of Machine Learning, Compiler Optimizations
Software Engineer
Citi BankPune, IndiaJuly 2023 – Jan. 2025
- •At Citi Bank, I wasn’t just writing code, I was collaborating with data scientists, experimenting with LLMs, and learning how to turn research into production systems.
- •Integrated LLMs (OpenAI GPT-3/4) for automated financial report generation and natural language analytics, reducing analyst workload.
- •Ensured production ML systems met enterprise security and compliance standards (JWT, SonarQube, Black Duck).
Software Engineer
Citi BankPune, IndiaJuly 2023 – Jan. 2025
- •At Citi Bank, I wasn’t just writing code, I was collaborating with data scientists, experimenting with LLMs, and learning how to turn research into production systems.
- •Integrated LLMs (OpenAI GPT-3/4) for automated financial report generation and natural language analytics, reducing analyst workload.
- •Ensured production ML systems met enterprise security and compliance standards (JWT, SonarQube, Black Duck).
Software Engineer
Citi BankPune, IndiaJuly 2023 – Jan. 2025
- •At Citi Bank, I wasn’t just writing code, I was collaborating with data scientists, experimenting with LLMs, and learning how to turn research into production systems.
- •Integrated LLMs (OpenAI GPT-3/4) for automated financial report generation and natural language analytics, reducing analyst workload.
- •Ensured production ML systems met enterprise security and compliance standards (JWT, SonarQube, Black Duck).
Software Engineer Intern
InterAct SoftwareMumbai, IndiaJun. 2022 – Aug. 2022
- •My first real taste of industry ML/AI! I loved seeing how research ideas could become features that real users rely on.
- •I developed RESTful APIs and backend services for ML-powered web apps.
- •Assisted in building NLP pipelines for user feedback analysis and sentiment classification (spaCy, NLTK).
- •Worked with product managers to translate ML research into production features, improving user engagement and retention.
Software Engineer Intern
InterAct SoftwareMumbai, IndiaJun. 2022 – Aug. 2022
- •My first real taste of industry ML/AI! I loved seeing how research ideas could become features that real users rely on.
- •I developed RESTful APIs and backend services for ML-powered web apps.
- •Assisted in building NLP pipelines for user feedback analysis and sentiment classification (spaCy, NLTK).
- •Worked with product managers to translate ML research into production features, improving user engagement and retention.
Software Engineer Intern
InterAct SoftwareMumbai, IndiaJun. 2022 – Aug. 2022
- •My first real taste of industry ML/AI! I loved seeing how research ideas could become features that real users rely on.
- •I developed RESTful APIs and backend services for ML-powered web apps.
- •Assisted in building NLP pipelines for user feedback analysis and sentiment classification (spaCy, NLTK).
- •Worked with product managers to translate ML research into production features, improving user engagement and retention.
B.Tech in ECE
Veermata Jijabai Technological Institute (VJTI)Mumbai, IndiaAug. 2019 – Jun. 2023
- •During my undergrad at VJTI, I got deeply involved in side projects and hackathons—building everything from speech recognition systems to web apps. This hands-on work sparked my passion for ML and set me on the path to research and engineering.
- •Graduated with distinction. Senior project: "Propoganda detection using BERT"
- •Relevant coursework: Probability & Statistics, Digital Signal Processing, Machine Learning.
B.Tech in ECE
Veermata Jijabai Technological Institute (VJTI)Mumbai, IndiaAug. 2019 – Jun. 2023
- •During my undergrad at VJTI, I got deeply involved in side projects and hackathons—building everything from speech recognition systems to web apps. This hands-on work sparked my passion for ML and set me on the path to research and engineering.
- •Graduated with distinction. Senior project: "Propoganda detection using BERT"
- •Relevant coursework: Probability & Statistics, Digital Signal Processing, Machine Learning.
B.Tech in ECE
Veermata Jijabai Technological Institute (VJTI)Mumbai, IndiaAug. 2019 – Jun. 2023
- •During my undergrad at VJTI, I got deeply involved in side projects and hackathons—building everything from speech recognition systems to web apps. This hands-on work sparked my passion for ML and set me on the path to research and engineering.
- •Graduated with distinction. Senior project: "Propoganda detection using BERT"
- •Relevant coursework: Probability & Statistics, Digital Signal Processing, Machine Learning.
About My Research Interests
My research interests lie in machine learning, deep learning, and natural language processing. I am particularly interested in building robust, scalable ML systems and exploring new algorithms for real-world impact. I have hands-on experience with research projects in NLP, computer vision, and large-scale data analysis.
Background
I have contributed to research and engineering projects at Citi Bank and during my graduate studies. My work spans deep learning, NLP, and data-driven systems. I am eager to collaborate on research that advances the state of the art in ML and AI.
I am currently seeking research assistantships, internships, or collaborations in the ML/AI domain.
Technical Skills
ML & Data Science
PyTorchTensorFlowscikit-learnNLPDeep Learning
Backend
PythonNode.jsJavaDjangoPostgresREST APIs
Tools
GitDockerAWSLinuxGIS
Frontend
ReactNext.jsTypeScriptTailwind CSS
Featured Projects
Here are some of the projects I've worked on
CGIT - Geospatial AI Query System
AI-powered tool that lets users ask natural language questions about map data (e.g., “Show me severe crashes in Fairfax County in the last 2 weeks”) and see results as map points, heatmaps, or spatial layers.
LLMsPostGISDjangoLeafletPython
Citi Bank
Full-stack application for financial data management and analysis, with ML-driven analytics and real-time dashboards.
ReactPythonD3.jsWebSocketsPrometheus
TerraMind: Natural Language Polygon Tool for GIS Developers
A tool for GIS developers that uses LLMs to convert natural language queries (e.g., “Give me the polygon for Central Park, NYC”) into precise area polygons, streamlining geospatial workflows.
LLMsPostGISPythonLeafletMapboxReact
Get In Touch
I'm always interested in hearing about new opportunities and exciting projects.
Feel free to reach out!
Phone
Location
Blacksburg, VA