AI Fullstack Developer at AGICO.AI (2025-06 – 2026-07)
Client Services at AGICO.AI (2026-01 – 2026-07)
- Reduced time taken from concept to contract by developing an internal prototyping tool that converts client specifications into detailed specs, screen designs and code artifacts.
- Enabled smoother collaboration and artifact sharing internally by migrating the internal tool from a local electron app to a cloud based system. Used pi coding agent as the agent harness and ddd-lite as an architectural guideline
- Prevented duplicate generation request and user drop off during long running AI generation by implementing resumable SSE stream. Providing users live progress updates that survive page refreshes and connection interruption.
Product (werd.ai) at AGICO.AI (2025-06 – 2025-12)
- Designed a multi agent workflow using plan and execute architecture in LangGraph. Enabling task decomposition for deep research.
- Enabled efficient agent debugging and tracing for multi agent architecture through integration with LangFuse.
- Reduced user wait times from 10mins (worst case) to 5 mins through parallelising the research phase of plan and execute architecture
AI-IoT (Intern) at GROUNDUP.AI (2024-09 – 2024-12)
- Assisted in developing and testing a workshop safety computer vision system, conducting offline deployments to collect data for training models, leading a 20% decrease in processing time and improving overall efficiency.
- Contributed to the development of an auto-annotation pipeline, utilizing the CVAT platform to refine annotations, and curating quality datasets for model training, resulting in a 30% reduction in time spent on manual data annotation
Fullstack Developer at GIGNITE.AI (2023-01 – 2024-09)
- Architected and implemented asynchronous request handling for distributed systems, including short polling and retry mechanisms with idempotent keys. Improved system robustness by 35%, increased successful request completion by 40%, and enhanced user experience through graceful failure handling.
- Pioneered platform observability improvements by integrating Axiom and Winston for comprehensive logging, enabling better monitoring and troubleshooting capabilities.
- Established feature flagging system using Growth Book, enhancing the ability to toggle and test new features dynamically across the platform.
- Enhanced platform stability and user experience by introducing skeleton loaders, fixing rendering issues, resulting in improved perceived loading times and reduced user frustration by 50%.
Fullstack Developer at XENDIT (2022-05 – 2024-01)
- Spearheaded spec-first REST API development, accelerating overall project timeline by 30% through enabling parallel backend and mobile development.
- Designed a Backend For Frontend (BFF) system using NodeJS, collaborating with mobile engineers to create a dynamic form builder; decreased mobile development time by 40% and halved partner integration efforts.
- Incorporated into a specialized payment team to implement a new payment channel, decreasing user transaction costs by 15% which enhanced the product competitiveness in the market.
- Developed React-based web platform for Know Your Customer (KYC) process, driving 23.3% of total KYC submissions and achieving a 58% completion rate, expanding user base without requiring app downloads.
Fullstack Developer (Intern) at REDDOT PAYMENT (2019-03 – 2019-10)
- Delivered a full-stack Internet Booking Engine (IBE) for hotels using VueJS and Express, serving both client-facing interfaces and internal tools, showcasing versatility in frontend and backend development.
- Led end-to-end project implementation, from gathering requirements with the sales team to designing and coding the platform, highlighting initiative and cross-functional collaboration skills.
- Constructed a comprehensive deployment pipeline, integrating GitLab CI/CD with AWS CloudFormation, enabling automated end-to-end deployment from code commit to cloud infrastructure provisioning, reducing deployment time by 70%.
- Directed seamless transition of database systems from DynamoDB to PostgreSQL, optimizing query response times by 35% and reducing technical debt, showcasing strategic problem-solving and flexibility to meet project demands.
