_.+._
(^\/^\/^)
\@*@*@/
{_S_C_}
I am a software engineer with a passion for astronomy. Seven years of professional experience across growth-engineering, retail, and enterprise SaaS. I own architecture, systems, and teams end-to-end.
I have grown up tinkering with computers and gadgets. I am also an amateur astronomer and run a small astronomy club. These days, I am mostly working on tinkering with frontier models and working on AI.
- Calendars team — building AI-powered agentic systems.
- Managed entire infrastructure — AWS, OpenSearch, MongoDB Atlas, Redis.
- Reduced p99 latencies on business-critical flows to sub 50ms.
- Solely designed and implemented security controls to achieve PCI-DSS and ISO certifications.
- Overhauled architecture to support 50% higher throughput on Notifications and Visitor Management System.
- Migrated x86 → ARM, reducing infra costs by 30%.
- Led and mentored a team of software engineers.
- Built habit trackers across three consumer brands — onboarded 500K+ users.
- Led an AI-powered doctor–patient call transcription system in 2021.
- Built a GenAI storybook platform for personalized children's bedtime stories.
- Engineered scalable contests with live leaderboards and automated prize distribution.
- Integrated CRED payments, Google Fit, LikeMinds, and OpenAI. Built consultation and instant appointment booking.
- Large-scale distributed retail POS for Dixons Carphone, UK & Ireland.
- Cross-functional teams across two timezones. AppDynamics monitoring during peak cycles.
- Wrote automation scripts to eliminate BAU manual tasks.
Building with AI since 2021 — before GPT-4, before the hype.
Building AI-powered agentic workflows for scheduling, calendar management, and autonomous task execution on production systems.
Led a team building an AI system to transcribe and structure doctor–patient calls. Applied speech-to-text and NLP to extract clinical insights from unstructured audio — before this was a solved problem.
Platform where parents generate personalised bedtime stories for their children. Handled prompt design, content moderation, and a backend that made generation feel instant.
Published research on clickbait detection using text classification. Early work that informs how I think about content quality and automated moderation pipelines.
- Go
- Node.js
- Rust
- PostgreSQL
- Qdrant
- ClickHouse
- AWS/GCP
- Docker
- OpenSearch
- Kubernetes
- Linux
- Grafana
- React
- Next.js
- Vue
- TypeScript
- Post-training
- Distillation
- Evals
- PyTorch
> Best ways to reach me: