Building systems that empower developers.
Hi, I'm Arun Kumar, a Software Developer (DevX) obsessed with API clarity, AI Native Workflows, and beautiful interfaces. B.Tech Electrical Engineering at NSUT - Delhi.
TypeScript
Flutter Expert
Cloud & Serverless
AI / MCP
01 // Experience
Technical Product Engineer Intern
Trelium AI (YC F25)- Monetisation Pipelines: Engineered end-to-end Monetisation pipelines, implementing token and workflow run steps-based billing logic that supports both individual credit pools and organization-level shared credit pools.
- Authentication & Security: Developed the core Authentication layer, integrating email/password flows, invite-only signups to prevent unauthorized access, and secure SSO logins for enterprise customers.
- Organization & Team Management: Architected comprehensive Organization and Team management features, establishing strict Role-Based Access Controls (RBAC) to ensure secure data isolation across workspaces.
- Super Admin Dashboard: Built a centralized Super Admin Dashboard to govern the platform, providing internal tools to seamlessly create, invite, and manage all underlying organizational entities from a single interface.
- Admin Tools: Deployed advanced Admin tools featuring User Impersonation and Multi-Client AI Chat, delivering a refined user onboarding experience and drastically accelerating customer issue resolution.
Software Developer (DevX) Intern
Digia Studio- Virtual Mobile Devices: Integrated virtual mobile devices directly onto the low-code builder dashboard canvas to mimic native emulators, achieving 100% UI fidelity testing and accelerating developer onboarding workflows.
- Custom UI Components: Published 10+ custom UI components to build the low-code builder interface, including masonry grids, paginated sliver lists, and a native chart widget, establishing predictable patterns for API consumers.
- Chart.js Integration: Authored and integrated the
chartjs_flutterabstraction layer natively into the core builder platform, bridging Chart.js to Flutter to drastically reduce data visualization implementation times. - Flutter/JS Bridge: Engineered a Flutter/JavaScript bridge allowing dynamic business logic injection without app recompilation, eliminating a major deployment friction point for third-party developers.
02 // Open Source Contributions
Active Contributor
Twenty: Modern Salesforce Alternative [40K+ GitHub Stars]- High-Impact Delivery: Merged 11 PRs across frontend and backend layers; received positive review comments from core maintainers Lucas Bordeau and Charles Bochet.
- [Feature Date Filtering]: Added
QUARTERas a new time unit for date filters, modifying TypeScript enums, regex parsing, UI dropdowns, and testing across 9 files to implement calendar-aligned logic. (Fixes #16674) - [GraphQL Filter Logic]: Extended
ACTORfield filtering to support source-based queries inturnRecordFilterIntoGql, implementing case-insensitive enum matching for operands. (Fixes #18099) - [Frontend State DX]: Resolved data staleness in virtualized tables by enforcing cache invalidation on metadata updates, preventing conflicting UI states across all enterprise workspace users. (Fixes #16388)
Contributor
CCExtractor- Memory Safety: Investigated and resolved a critical heap OOB buffer over-read in MP4 atom parsing, significantly improving memory safety for downstream data pipeline consumers. (Fixes #2179)
Published Package (Sole Author)
chartjs_flutter- Chart.js Abstraction: Authored a core abstraction layer binding Chart.js natively to Flutter; actively maintained with clear documentation to reduce time-to-implementation for frontend data visualization consumers.
03 // Full Stack Projects
FaceSwapper AI
Generative AI Pipeline & Editor- Video Processing & Auth: Architected a full-stack video processing platform using Next.js 15 Server Actions, integrating Better-Auth for secure access and Resend to automate curated user onboarding and verification email sequences.
- Rate Limiting & Job Orchestration: Designed a Redis sliding-window rate limiter to strictly enforce AI generation quotas, coupled with a custom asynchronous state machine to systematically track and resolve external AI job polling.
- Storage Pipeline: Built a robust storage pipeline via AWS SDK and Cloudflare R2 utilizing presigned URLs for uploads, and engineered a sequential deletion workflow for assets and thumbnails to prevent orphaned elements.
- Expense Analysis via AI: Deployed Groq API (Llama-3-70B) to dynamically analyze user spending habits, automatically generating tailored monthly expense distributions to help users proactively manage and control their budgets.
- Background Tracking: Engineered a background tracking engine that detects transactions via SMS, push notifications, or on-screen node capture as a fallback, instantly triggering a localized popup for seamless expense logging.
- Realtime Backend & Visualization: Built a Supabase Realtime backend integrated with dynamic visual charts, enabling users to consistently record daily transactions, track live financial trends, and monitor their long-term wealth goals.