Harihara Roopan

I’m an SDE 2 with 3 years of experience building and owning production systems, mainly across Backend Engineering, Node.js, MERN, AWS and distributed systems.

In my current organisation, I took end-to-end ownership of a 0-to-1 eSign product, working across backend APIs, third-party integrations, system reliability, performance optimisation and customer-facing features. I’ve worked closely with external vendors, handling different vendor APIs and behaviours under the hood while exposing a consistent experience to customers.

I enjoy solving problems beyond just building CRUD APIs. I like understanding the actual problem, thinking from first principles, finding bottlenecks and improving the system for reliability, scalability and performance. I’ve worked with AWS, Node.js, React, TypeScript, MongoDB, Redis, Elasticsearch, Docker and REST APIs, and have experience working with production systems and external integrations.

I’m currently particularly interested in Backend Engineering, Distributed Systems, System Design, Cloud Infrastructure, Kubernetes and AI/LLM Infrastructure. I’m also exploring LLM Engineering, AI agents, model serving and production AI systems.

I enjoy 0-to-1 products, high-ownership environments and problems where there is a lot of room to improve the product and the underlying engineering.

Software Development Engineer 2 - Hyperverge Technologies - Bengaluru, Karnataka

(2026-05)

• Completed eSign frontend refactoring by introducing code-splitting and dependency optimizations, reducing JavaScript bundle size by 60% and enabling users on low-bandwidth (2G) networks to complete eSign journeys.
• Proactively built and shipped an AI copilot POC that turns natural-language prompts into validated low-code API configurations — combining retrieval-augmented generation (RAG), a self-validation and repair loop, and schema drift detection — cutting API build time by 80% (4 hours → 1 hour).
• Spearheaded the rollout of a centralized LLM Gateway with fallback, failover, and real-time monitoring across 20+ services; automated failovers shield ∼0.4% of monthly LLM traffic from provider degradation, and anomaly-detection MTTR dropped from days to minutes (>99% faster).
• Maintain end-to-end reliability of the eSign platform — React frontend, Node.js backend, and a PDF-intensive processing service handling 13,000+ signatures daily.

Software Development Engineer 1 - Hyperverge Technologies - Bengaluru, Karnataka

(2024-06 - 2026-04)

• Built a self-serve eSign dashboard with Role-Based Access Control (RBAC), enabling management of 30+ templates by non-engineering teams and reducing client onboarding time from 5 to 2 days for 20+ enterprise clients.
• Implemented PII data encryption and client-level data isolation as part of ASP compliance requirements, supporting production onboarding of multiple fintech and NBFC clients onto the eSign platform.
• Led a quarter-long conversion initiative for a top wealth-management client: built funnel analytics from vendor-pushed logs to pinpoint drop-offs, drove vendor-side UI fixes, and shipped Metabase/Hotjar-informed UX improvements — lifting journey conversion from 93% to 96% (∼40% fewer drop-offs).
• Integrated biometric signing mode into multi-signer flows and drove eStamping to production readiness through active vendor coordination; implemented ESP fallbacks with health-based vendor routing, maintaining 99.95% platform uptime through vendor downtimes.
• Drove API optimizations across the PDF generation module (text-size algorithm, input parser), sign URL creation, and callback handling, reducing P90 latency by more than 50% end-to-end.
• Reduced Time-to-Detection (TTD) in the alerting system by 99% (30 min → 1 min) by optimizing monitoring workflows.

Software Development Engineer Intern - Hyperverge Technologies - Bengaluru, Karnataka

(2023-07 - 2024-05)

• Automated the development of proof-of-concept (POC) results by creating an API request simulator tool, resulting in expedited responses to new clients and increasing customer conversion rates by 20%.
• Executed a zero-downtime embedding-model migration for a face-match service handling 1M+ calls/day — ran dual-writes to parallel Milvus vector collections, backfilled existing enrollments, and phased the client cutover — improving match accuracy with zero production degradation.

Bachelor of Technology - Information Technology - Kumaraguru College of Technology (2020-10 - 2024-04)