about

I'm a 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, software engineering, and AI product development, including building optimized backend systems, developing deep learning models for satellite imagery, and creating AI solutions for content moderation and real-time accident detection.

My work has been presented at esteemed venues like EMNLP, ICML, NeurIPS, MIT and Stanford's TSRC. For any research opportunities, please reach out to me via email or github.

work history

Lead Engineer for Arbiter, a social media analytics and fact checking tool, processing over 750M+ data points across social media platforms.
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 80,000+ Instagram posts with real-time transcription and fact-checking pipelines.
Built a deep-learning-driven satellite super-resolution system using PyTorch, SRCNN, and GANs, achieving 96% SSIM for high-fidelity remote-sensing imagery.
AI Agents for mock tests and revision built by fine-tuning DeepSeek using Generative Preference Optimization (GPO).
Worked with multiple MNCs and enterprises to integrate AI workflows into production systems as a freelance software engineer.
Engineered a multimodal VLM pipeline for real-time accident detection using BLIP-2, LLaVA-1.5, TensorRT, and RAG-enhanced contextual intelligence.
Developed a large-scale AI interviewing and resume-analysis system using GPT-4 Turbo, PySpark, and Snowflake, with a full cloud migration to GCP for improved scalability and efficiency.