← back

about

I'm a Founding Machine Learning Engineer at SimPPL, focused on building scalable AI systems with real-world impact. I also enjoy building with LLMs in my free time and learning more about ML, RL, DL, LLMs and their theory.

I've worked across research and software engineering, including building optimized backend systems, shipping healthcare AI applications, and creating AI solutions for content moderation and social media analysis.

My work has been presented at esteemed venues like NeurIPS, ICML, AAAI, EMNLP, TrustCon and Stanford's TSRC, and has been supported by over $2.5 million USD in research grants from institutions including Columbia University, Harvard AI Venture Lab, INSEAD, Deutsche Welle, Google exploreCSR Award, and MIT Delta Accelerator. For any research or work opportunities, please reach out to me via email or github.

OctNovDecJanFebMarAprMayJunJulAugSep
19033 contributions in the last year
LessMore

work history

  • Lead ML Engineer for Arbiter, a social media analytics and fact-checking platform, owning the backend and AI systems that ingest 50M+ posts weekly, with 2B+ monthly views across the content we audit.
  • Deployed Arbiter across 20 countries, used by 150+ organizations including Deutsche Welle, INSEAD, NEST Mongolia, and Jagran Media in India.
  • Engineered large-scale distributed systems for a fintech platform, implementing auto-scaling Kubernetes workloads, CI/CD automation, and real-time observability for resilient high-traffic operations.
  • Built InfluenceCheck, a large-scale social media authenticity engine processing 5 million+ Instagram posts with real-time transcription and fact-checking pipelines.
  • Engineering Sakhi, an MIT-incubated multilingual WhatsApp chatbot delivering expert-verified maternal health information to 44M+ Bengalis across Bangladesh and India.
  • Founding Intern at EvaluAItor, leading the development of the DSA practice platform and building the execution engine on Judge0.
  • Fine-tuned Llama 2 70B with QLoRA on DSA and system-design assessment data for adaptive mock tests and personalized revision, then served it through a SageMaker endpoint integrated with an autoscaling EC2-backed ECS service.
  • Worked as a freelance software engineer building applications and integrating AI workflows for enterprises across Asia, Europe, and Australia.