Venkatesh Prasad Ravichandran
I don't wait for the right problem.
I work with the one in front of me.
- Legal AI with near-zero hallucinations.
- A roofing company's inbound workflow, automated end-to-end.
- A graduate complexity theory course, co-designed and shipped.
MS in Business Analytics & AI at UT Dallas · Dean's Excellence Scholar · Already shipping.
The most interesting work lives behind a firewall. Ask me about it.
70%
less manual triage
Coalition Restoration Roofing - email classification replaced a full daily sorting workflow (~40 inbound emails/day).
40%
admin effort reduced
Coalition Restoration Roofing - workflow automation across homeowners, roofers, and insurance agents.
60%
traffic growth
ArtSciLab + OC4ES - information architecture overhaul and migration from DigitalOcean to AWS.
~80%
stylistic alignment
Fred the Heretic - RAG output evaluated by domain experts against Prof. Fredric Jameson's writing corpus.
How I Handle Complexity
Behavior before buzzwords.
AI
I build around the workflow, not the model. RAG when retrieval matters. Agents when the task has steps. Classification when speed beats accuracy.
Product
I do the messy work before the roadmap: the user calls that reframe the problem, the PRD that defines done, and the stakeholder conversations that kill weak ideas early. Scope, alignment, and shipping discipline are built in from day one.
Data
Bad data is a product problem, not just a pipeline problem. I instrument KPIs, fix CRM chaos, and make sure dashboard numbers reflect real operations so teams can act on them with confidence.
How I Think
The operating system behind the work.
Most people pick a lane. I pick the problem.
I have built in a law-adjacent research lab where being wrong had professional consequences, and in a field operations business where users were contractors on rooftops. Different domains, same job: find what is broken, understand why, and ship the fix.
Volume. Variety. Velocity.
I context-switch without losing precision and ship in environments where requirements move fast. In the last 12 months I shipped legal AI, roofing automation, and a hackathon product across different stacks and problem domains. That is not luck. It is a repeatable operating system for execution under constraint.
Where I've Shipped
AI & Automation Intern
Coalition Restoration- 40% administrative effort reduction by redesigning customer communication and internal workflow automation end-to-end.
- Spearheaded infrastructure migration (GoDaddy → Azure), improving platform reliability by 25% and reducing annual hosting costs by 15%.
- Shipped an automated email classification and prioritization system, cutting manual triage time by 70%.
- Drove CRM integration strategy (JobNimbus + CompanyCam), eliminating duplicate data entry and establishing a single source of truth for operational data.
Research & Product Analyst
ArtSciLab — UT Dallas- ~80% stylistic alignment achieved by co-owning product framing and delivery of the Fred the Heretic RAG system.
- Led platform reliability and information architecture improvements across ArtSciLab and OC4ES web, increasing user traffic by 60%.
- Executed server and domain migration from DigitalOcean to AWS, reducing downtime by 35% and enabling future scalability.
Graduate Trainee
Wisdom Schema- 100% downstream data accuracy restored at Wisdom Schema, an aviation data analytics firm, by resolving systemic time-zone inconsistencies in U.S. air-traffic datasets.
- Partnered with analytics and engineering teams to align data outputs with operational decision-making needs.
Built on Strong Fundamentals
M.S. Business Analytics & Artificial Intelligence
The University of Texas at Dallas
B.Tech, Computer Science
SASTRA University
Skills & Tech Stack
Tools I use to move from ambiguity to delivery.
Product Management
Data & Analytics
Platforms & Systems
Selected Projects
AI, product, and data work in practice.
Fred the Heretic
Graph-Augmented AI Product
This is what happens when RAG is taken seriously enough to measure stylistic fidelity against a literary corpus.
Agentic AI Career Advisor
Multi-Agent Workflow Product — Hackathon Winner
Built and demoed in under 48 hours, won, then expanded into a stronger multi-agent guidance workflow.
Coalition Operations Intelligence
Roofing Ops Automation Platform
Runs in production daily across CRM workflows, automation, and KPI tracking. The demo is private; the architecture is not.
AI Judge
Legal Reasoning System
Near-zero hallucination legal AI with verified citations, built for workflows where being wrong has consequences.
What I Can Ship in 30 Days
Practical outcomes in the first month.
- Map manual workflows and rank them by automation ROI.
- Instrument the KPI the team debates but cannot measure consistently.
- Ship a working prototype of the capability that has been stuck in planning.
Leadership & Achievements
Third-party proof from competitive environments.
Agentic AI Hackathon Winner — UT Dallas
Led a practical agentic AI product from idea to demo under time pressure.
Dean’s Excellence Scholar
Awarded merit-based scholarship at The University of Texas at Dallas for the M.S. Business Analytics program.