Open to full-time roles
Hi, my name is
Mehul Mangave.
Building polished full-stack products, end to end.
Software Engineer with 2+ years of production experience at State Street and Syngenta, focused on building full-stack web applications end to end. I develop responsive React, Next.js, and TypeScript front-ends wired to REST APIs and relational databases, and deploy them on AWS with Docker and CI/CD. My work also carries a strong cloud and applied-AI thread — from a production RAG assistant to an award-winning multi-agent dashboard, across projects like SnapTrip and QueueFlow. Actively seeking full-time Software Engineering roles.
97
Lighthouse performance score on the SnapTrip web app
10K+
Users served by production platforms I've built & maintained
60%+
Manual effort eliminated through automation & internal tooling
01. Professional Experience
Software Engineering Intern
State Street Bank and Trust
June 2025 – Aug 2025 Boston, US
- •Developed an internal full-stack knowledge assistant using React, TypeScript, Python, and REST APIs, enabling engineers to search Confluence documentation and receive context-grounded responses through a responsive web interface.
- •Designed backend services for document ingestion, vector retrieval, authentication, request validation, and feedback collection, improving application reliability and reducing repeated platform-support requests.
- •Built an administrative dashboard for monitoring document synchronization, user queries, response quality, and system health, and deployed the application through a containerized CI/CD workflow on AWS.
Software Engineer
Syngenta
June 2023 – July 2024 Pune, India
- •Led the modernization of a legacy operations portal using React, TypeScript, and Next.js, translating high-fidelity HTML/CSS designs into reusable production components and reducing new-feature development time by 35%.
- •Built responsive dashboards and workflows for resource requests, deployment approvals, and cost visibility, integrating frontend interfaces with REST APIs and supporting more than 10,000 users across desktop, tablet, and mobile devices.
- •Created a shared component library and frontend design system with standardized forms, tables, navigation, validation, and accessibility patterns, reducing duplicated UI code by 40% and improving consistency across internal applications.
- •Improved application performance through code splitting, lazy loading, client-side caching, and bundle optimization, cutting initial page-load time by 45% while owning features through testing, code review, release, and production troubleshooting.
Software Engineering Intern
Syngenta
Aug 2022 – Jun 2023 Pune, India
- •Built Node.js REST APIs for application configuration, deployment requests, and operational data, implementing request validation, structured error handling, authentication, and integrations with AWS services.
- •Developed responsive internal web applications using React, TypeScript, HTML5, and CSS3, converting business requirements and UI designs into reusable components used by engineering and operations teams.
- •Added unit and integration tests, resolved defects across the frontend and backend, and automated application builds and deployments using Git, Docker, and CI/CD pipelines, reducing release time and improving production reliability.
Open-Source Terraform Modules
— public proof of work02. Technical Projects
1st Prize — Most Innovative Use of Agents · Harness Engineering Hack
Magpie — Ambient Study Agent
Hackathon · 2026
Guild.ai AgentsClickHouseNext.jsOpenUIPython
- ▹Built an ambient study tool that watches the system clipboard and treats every copy as an implicit highlight — snippets are classified into topic notebooks, streamed to ClickHouse, and surfaced on a live dashboard with summaries and a chat assistant.
- ▹Designed the system as four composed, single-purpose Guild.ai agents (classify, summarize, re-file, chat) kept pure as text-in/text-out, with all database access isolated to the watcher and API routes.
- ▹Hardened the demo with a non-blocking capture queue, server-side validation and deterministic fallback for malformed LLM output, and scoped ALTER TABLE mutations so the dashboard never breaks on a bad agent response.
SnapTrip: AI-Powered Travel Planning Application
Aug 2025 – Dec 2025
ReactJavaScriptMaterial UIFramer MotionOpenAI API
- ▹Developed a responsive, multi-step travel-planning application using React, JavaScript, Material UI, and Framer Motion, transforming destination, schedule, interests, pace, and accessibility inputs into a polished experience across desktop and mobile devices.
- ▹Integrated the React frontend with a backend REST API that sends structured user preferences to the OpenAI API and processes JSON responses to generate and regenerate personalized multi-day itineraries.
- ▹Implemented client-side trip persistence using localStorage, PDF itinerary export with jsPDF, and light/dark themes, while achieving a Lighthouse performance score of 97 and accessibility score of 87.
QueueFlow: Appointment & Waitlist Management Platform
Jan 2026 – Mar 2026
TypeScriptNext.jsNode.jsPostgreSQLReact
- ▹Developed a full-stack appointment and waitlist management platform using TypeScript, Next.js, Node.js, and PostgreSQL, supporting role-based workflows for customers, employees, and administrators.
- ▹Designed REST APIs and relational database models for appointments, employee availability, customer records, and waitlists, including input validation and scheduling-conflict detection.
- ▹Built reusable, responsive React components and an administrative analytics dashboard for tracking bookings, wait times, and staff utilization across desktop and mobile devices.
Dynamic Weight Learning for Self-RAG
Aug 2025 – Dec 2025 · Team of 2
PythonPyTorchLLMRAG
- ▹Built a dynamic weighting framework for Self-RAG pipelines that re-ranks and adaptively weights retrieved passages by query relevance, replacing static weighting that treated all retrieved documents equally regardless of context.
- ▹Implemented adaptive scoring in Python and PyTorch across 1,000+ query evaluations, improving response accuracy and relevance by ~18% while keeping inference overhead low through optimized retrieval scoring logic.
Omnireasoner: Knowledge Graph Reasoning System
Aug 2025 – Mar 2026 · Team of 4
Knowledge GraphsLLMsGraph Traversal
- ▹Built a knowledge-graph-augmented reasoning pipeline combining graph traversal with LLM prompting over a 10K+ entity graph, improving factual consistency on complex multi-hop question answering.
- ▹Generated interpretable reasoning paths through entity linking and graph traversal, reducing hallucinated responses by ~15% in evaluation by grounding answers in retrieved graph relations rather than model priors.
Driver Assistance System
Jan 2022 – Jun 2022
SIFTPCAMachine LearningPython
- ▹Developed a system for autonomous vehicles that achieved 95% accuracy in predicting the turning direction of the vehicle ahead. Enhanced driver awareness by providing real-time audio feedback, reducing reaction time by 30%.
- ▹Utilized SIFT feature extractor for feature extraction, Principal Component Analysis for dimensionality reduction, and 6 types of machine learning classifiers for classifying the images.
Arrhythmia Detection via SVM
Aug 2021 – Dec 2021
ECG Signal ProcessingSVMScikit-learn
- ▹Mitigated heart attack risk by early detection of arrhythmia through analyzing processed ECG signals using classical ML techniques to accurately classify arrhythmia types.
- ▹Presented at the 4th International Conference on Robotics, Intelligent Automation and Control Technologies (VIT Chennai in association with Teesside University and Bochum University).
03. Technical Mastery
Frontend
TypeScriptJavaScriptReactNext.jsHTML5CSS3Tailwind CSSMaterial UIFramer Motion
Backend & APIs
Node.jsREST APIsPythonDjangoOpenAI API
Databases
PostgreSQLSQLMongoDBDynamoDBNoSQL
Cloud & DevOps
AWSAzureDockerKubernetesTerraformCI/CDGitGitHub Actions
04. Education
San Jose State University
Master of Science in Computer Engineering
May 2026 | San Jose, CA, USA
Vishwakarma Institute of Technology
Bachelor of Technology in Electronics and Telecommunications Engineering
May 2023 | Pune, India
Relevant Coursework: Designing Data Structures, Algorithm Design, Operating Systems, Advanced Computer Design, Systems Software, Artificial Intelligence and Data Engineering
05. Get In Touch
I am actively seeking full-time Software Engineering roles. Whether you have an opening or just want to connect, my inbox is open.
+1 (408) 210-8953San Jose, CA