Independent Applied AI Practice
I work independently across a broad portfolio, designing and shipping applied AI products, research infrastructure, developer tools, and human-centered systems.
Recent work spans coastal field intelligence, agent tooling, real-time movement coaching, evidence-grounded retrieval, digital humanities, and computer vision for public health.
My background includes Research Data Scientist at Microsoft, applied AI work at DeepIntent, collaborations with university research teams, and Principal AI Scientist & Founder at Byteflows.
Founders, research groups, and organizations looking to build an AI product, workflow, or discovery system can reach out directly.
Contact
Independent products, open-source tools, consulting work, and research collaborations.
Focused career platform for people building, deploying, and leading AI. Role Atlas indexes opportunities across emerging startups and established teams, verifies availability, and links directly to company application pages.
Live platform
atlas.byteflows.com
Mobile-first field-intelligence proof of concept for San Juan, Puerto Rico. Onda combines National Weather Service surf-zone forecasts, recent NOAA buoy observations, and fast field logging in one analysis-ready workflow.
Package manager for repository-scoped agent tooling. Nodus installs skills, agents, rules, commands, and MCP settings from versioned sources, then writes the runtime files each supported coding agent actually uses.
Browser-based proof of concept for real-time movement coaching. Nabi guides a structured arm set, measures range, symmetry, stability, control, and speed, and keeps camera frames and movement analysis on the user’s device.
Document intelligence workspace using RAG and hybrid retrieval to explore large research corpora with transparent evidence and citations. Built for exploration without losing the underlying evidence.
Live URL (NYU VPN or on-prem required)
research-rag-workbench-git-rts-genai-research-hub-dev.apps.cloud.rt.nyu.edu
AI-assisted exploration of early English books using translation, semantic indexing, and retrieval to support digital humanities research. Extends the portfolio into humanities-focused corpus analysis.
Live URL (NYU VPN or on-prem required)
early-books-rag-workbench-rts-genai-research-hub-dev.apps.cloud.rt.nyu.edu
Platform mapping generative AI projects, use cases, and capabilities across NYU research groups. Maps active generative AI work across NYU and makes it easier to discover, compare, and understand.
Access
Request early access
Hands-on workshops helping faculty and staff use generative AI tools for research, teaching, and workflow automation. Connects AI capability to day-to-day institutional use.
Reference URL
wp.nyu.edu/fas-edtech/flash-forc
Curated AI-assisted clinical support prototype focused on substance use, overdose prevention, harm reduction, and culturally responsive communication. Designed as a T3/T4 translational research pilot to help frontline clinicians access trustworthy, non-stigmatizing, evidence-grounded guidance in real-world care settings.
Research team
Luther Elliott, NYU School of Global Public Health - Associate Research
Scientist
Andi Shehu (Applied AI Specialist)
POC URL (NYU VPN or on-prem required)
rts-genai-rag-apps-dev.apps.cloud.rt.nyu.edu
Agentic digital marketing workflows for turning public content, events, and audience intent into updated discovery experiences, campaign pages, and conversion paths. Designed for outreach, audience routing, and lightweight content automation across public-facing digital experiences.
Access
Request early access
Multimodal system detecting and classifying tobacco products from retail shelf images for policy and public health analysis. Applies computer vision to a policy-relevant public health problem.
Status
Private previewA compact view of the experience behind the work.
Experience
10+ years across Microsoft, DeepIntent, Byteflows, consulting, and research collaborations.
Domains
Healthcare, advertising technology, research, finance, public-interest systems, and developer tooling.
Training
PhD in Physics, focused on quantum information theory, and an MS in Mathematics.
Executive communication, institutional analysis, and applied AI strategy work.
Executive deck synthesizing researcher needs, peer platform choices, and recommended next steps for institutional AI support. Developed in an NYU context, but representative of broader higher-ed AI strategy and advisory work.
Status
Private previewEvent and community work in New York.
DataBarOnTheRun was a New York series I ran for people working across data, analytics, AI, startups, and business. Some nights were more social and some were closer to roundtables, but the format stayed small so people could talk for more than five minutes.
Public URL
andishehu.byteflows.com/databarontherun/For research collaborations, platform design, or applied AI consulting, get in touch.