Cancer Biology & Hematology
Investigating disease mechanisms in leukemia, fibrosis, and inflammation. Focusing on apoptosis, cell cycle regulation, and cellular differentiation in hematopoietic systems.
About Me
Bridging AI and Biomedicine
I am an Assistant Professor at the Icahn School of Medicine at Mount Sinai in New York City, specializing in Medicine, Hematology, and Medical Oncology. My work stands at the forefront of translational research, applying cutting-edge artificial intelligence to solve complex biological problems.
My research focuses on dissecting the mechanisms of cancer biology, hematopoiesis, and immunology using single-cell multiomics and machine learning. By developing evidence-based AI systems, I aim to bridge the gap between computational discovery and clinical application, creating tools that are not only statistically rigorous but also interpretable for clinicians.
As the architect of the Universal Skill Definition Language (USDL), I am pioneering a standardized framework that allows biomedical AI agents to operate across platforms—scaling scientific analysis from the bench to the bedside.
Biomedical Domain Expert
Hematology, Oncology, Immunology & Stem Cell Biology
AI Systems Architect
Clinical Decision Support & Agentic Workflows
Multiomics Analyst
Single-cell RNA-seq, Spatial Transcriptomics & Genomics
Quick Facts
- Position: Assistant Professor
- Institution: Mount Sinai, NYC
- Department: Medicine, Hematology & Oncology
- Focus: AI, Cancer Biology, Genomics
- Email: md.babu.mia@mssm.edu
Research
Research Interests & Focus Areas
Clinical AI & Decision Support
Developing AI agents for clinical trial matching and evidence-based recommendations. Systems designed with uncertainty quantification and audit-ready provenance for healthcare.
Drug Discovery & Repurposing
AI-powered compound screening and target identification pipelines. Utilizing multi-agent orchestration to accelerate the discovery of therapeutics for complex diseases.
Single-Cell & Spatial Omics
Complete analysis stacks for scRNA-seq and spatial transcriptomics. Automating cell type annotation and spatial analysis to uncover cellular heterogeneity.
LLM Agents & USDL
Pioneering the Universal Skill Definition Language (USDL) to standardize biomedical AI capabilities across LLM platforms (Claude, ChatGPT, Gemini) for reproducible science.
AI Governance & Safety
Building robust governance frameworks with model registries, drift monitoring, and clinical safety gates to ensure responsible AI deployment in patient care.
Projects
Featured Open Source Work
Explore my contributions to biomedical AI and computational biology
Featured
8
LLMs-Universal-Life-Science-and-Clinical-Skills
A comprehensive platform for biomedical AI skills using the Universal Skill Definition Language (USDL). Enables cross-platform deployment of clinical decision support, drug discovery, and genomics workflows.
Jupyter Notebook
Python
YAML
2
cellidentifierdx
High-performance Python library for automated cell type annotation in scRNA-seq workflows. Integrates curated marker databases with statistical validation for precise cellular identification.
Python
scRNA-seq
ML
1
DNAdamageclassifierDx
Machine learning framework for classifying DNA damage patterns from genomic data. Utilizes ensemble methods to provide interpretable feature importance and high-accuracy classification.
Python
ML
Genomics
1
ncbi-geo-pubmed-search
Advanced programmatic interface for querying NCBI GEO and PubMed. Streamlines meta-analysis by enabling complex filtering and automated retrieval of biomedical datasets.
Python
API
NCBI
Platform
BioMedAI Platform Components
Production-ready tools engineered for evidence-first biomedical teams
Data Analysis Skill
Unified Analytics Workbench
A full-stack analytics agent bundling Python (Pandas, NumPy), R, SQL, Tableau, and Power BI so translational teams can jump from exploratory data analysis to executive storytelling without context switching.
- Drop-in USDL module for LangChain, AutoGen, Semantic Kernel
- Embedded data quality gates (missingness, MAD-based outlier checks)
- Reusable YAML skill definition with citations for every workflow
Clinical Launchpad
TrialGPT + Clinical Notes
Precision trial matching and SOAP note structuring agents that transform EMR exhaust into decisions clinicians can trust. Each recommendation bundles rationale, contraindication flags, and audit-ready provenance.
- Optimized for NIH datasets and Mount Sinai clinical protocols
- Escalation states for insufficient data + uncertainty overlays
- Plug-and-play with SMART-on-FHIR roadmaps
Research Ecosystems
scverse + Seurat Integration
Complete single-cell analysis stacks cloned locally for rapid STAgent/BioMaster prototyping, including annotated tutorials, benchmarking notebooks, and reproducible pipelines.
- Spatial transcriptomics orchestration with STAgent + BioMaster
- Ready-to-run MCP servers for PubMed & ClinicalTrials retrieval
- Versioned playbooks for ICP, CRISPR, and cell annotation tasks
Publications
Publications
For my complete, up-to-date list of peer-reviewed work, please visit my Google Scholar and ORCID profiles.
Skills
Technical Expertise
Programming Languages
AI & Machine Learning
Bioinformatics
Tools & Platforms
Experience
Career & Education
Present
Assistant Professor
Icahn School of Medicine at Mount Sinai
Leading research in the Department of Medicine, Hematology, and Medical Oncology. Developing AI-driven solutions for cancer biology, single-cell multiomics, and clinical decision support systems.
Research
Artificial Intelligence Group
Mount Sinai Health System
Building production-ready clinical AI systems with governance frameworks, evidence-backed workflows, and multi-agent orchestration capabilities.
Education
Indiana University
PhD
Doctoral research focusing on Computational Biology and Bioinformatics.
Roadmap
BioMedAI Platform Roadmap
Phase 1
Trust + Workflow
Clinical report objects, provenance logging, and error-mode UX so every output is chart-ready.
- PredictionCard schema with uncertainty + contraindication flags
- Evidence packs referencing internal docs and guidelines
- Escalation states ("needs expert review", "insufficient data")
In Progress
Phase 2
Connectors + Cohorts
Data flows from source systems with validation, cohort builders, and monitoring dashboards.
- FHIR + CSV ingestion adapters with unit/ontology normalization
- Cohort builder with provenance-aware exports
- Calibration, drift, and subgroup safety monitors
Planned
Phase 3
Multi-modal + Knowledge
Document ingestion, knowledge graph services, and plugins for imaging, pathology, and genomics.
- Governed retrieval store with citations + redaction
- Optional biomedical knowledge graph for agents
- Modular multi-modal adapters (DICOM, omics, notes)
Future
Contact
Let's Connect
Interested in collaboration, research opportunities, or just want to chat?
Location
Manhattan, New York City, USA
Institution
Icahn School of Medicine at Mount Sinai