Work / 01
Selected projects &
professional experience.
A curated set of work spanning healthcare AI, multi-agent systems, generative models, and applied research. From startups I co-founded to research labs I contributed to.
IonTheFold
June 2025 to Present- Co-founded a research-focused startup reimagining protein and antibody design with AI, unlocking breakthroughs for drug discovery and biotechnology.
- Led development of deep learning pipelines with the Grinter Lab (University of Melbourne) and Knott Lab (Monash University), leveraging NGS-derived sequence datasets.
- Designed a fusion architecture combining ESM-2, APBS, and ProteinMPNN with a novel electrostatic-aware loss function, achieving 6.53% improvement on charged proteins and 4.14% ± 0.026 overall gain in sequence recovery.
Quelea
June 2025 to Present- Developed a multi-agent video analysis platform with real-time audio, visual, sentiment, and multimodal intelligence; integrated agents for ASMR detection, beauty scoring, pose analysis, OCR, and virality prediction.
- Built a custom model inspired by NexTGPT trained on 5,058 influencer videos to achieve state-of-the-art results in multimodal understanding; preparing for arXiv publication.
- Finetuned and deployed Qwen 2.5 Vision Instruct via QLoRA on Runpod, boosting overlay text extraction accuracy in diverse video conditions.
University of Arizona
March 2025 to Present- Developed digital phenotyping pipelines integrating MindLamp and deploying RADAR-Base on AWS EC2 for continuous ingestion of mobile sensor data, supporting MyDataHelps deployment for a postpartum health study.
- Engineered a multi-agent architecture unifying wearable and phone sensor data by building knowledge graphs, KNN clusters, and text embeddings.
Lumenci
January 2024 to December 2024- Developed and deployed an automated portfolio analysis system using ML algorithms and advanced feature engineering, reducing Associate consultants' work from 40-50 hours per project to minutes.
- Accelerated portfolio analysis by 82% by leveraging GCP's GPU for backend operations.
- Developed an AI workbench using CrewAI and OpenAI agents in a multi-agent framework, improving operational efficiency by 70% and boosting model accuracy from 6% to 72%.
Indian Institute of Technology, Indore
January 2022 to May 2022- Developed a Graph Neural Network architecture from scratch to process and train on a decade's worth of image data from NASA's SOHO satellite, achieving 91% accuracy in time series prediction.
- Led a team of 3 researchers experimenting with CNN, RNN, GAN, and GNN for time series forecasting and image classification tasks.
FeaturedShifaMind
Explainable healthcare AI capstone with diagnostic reasoning grounded in medical evidence.
FeaturedIonTheFold
Protein & antibody design with electrostatic-aware fusion architecture.
FeaturedCOMET-SEE
Comet detector trained on NASA SOHO data, deployed on Hugging Face Spaces.
FeaturedJobHuntAI
Multi-agent job search assistant that scrapes, scores, and applies on your behalf.
BioSphereAI
Hackathon-winning bioinformatics platform for environmental data.
Salsabil
AI-powered productivity & wellbeing companion.
VitaShifa
Personal health AI prototype merging RAG and structured medical knowledge.
Protein Protein Interaction GNN
Graph neural network for predicting PPI from sequence and structure.
Afny AI
Conversational AI assistant deployed for community use.
AutoPharma
Pharmaceutical assistant powered by multi-agent reasoning.
Image Detection
GAN + Vision Transformer based AI-generated image identifier.
Jarvis AI
Personal voice-driven AI assistant with multi-modal capabilities.
Solar Weather Forecast
Time-series forecasting of solar storms using NASA SOHO data.
Archer · Model Rocketry
Patented self-landing model rocket flight computer.
Loan Default Prediction
End-to-end ML pipeline for German bank loan default prediction.