KamyavardhanDave
"Aut inveniam viam aut faciam."
DoctorCopilot
OCR -> extraction -> signal analysis -> doctor workspace
A product that turns uploaded medical reports into readable case context, trend history, and role-specific views for doctors, patients, and admins.
Built with: FastAPI · Next.js · OCR · OpenAI · Railway
Product flow
Live railUpload report or CBC panel
OCR and field extraction normalize the source
Signal interpretation produces structured findings
Doctor workspace assembles the case context
Role snapshot
Doctor
Case review
Patient
Trend follow-up
Admin
Operations
Deployment
Frontend
Vercel-hosted product shell
Backend
Railway-hosted API and report processing
Clinical Validation Pipeline
Brain Tumor AI
MRI -> model comparison -> uncertainty -> Grad-CAM explanation
An MRI demo that compares two models, shows uncertainty, and adds Grad-CAM so the result can be checked instead of just accepted.
Built with: PyTorch · Streamlit · EfficientNet · Grad-CAM
Why it matters
Accessibility Product Concept
Lexara Air
Reading support -> visual scaffolding -> calmer interface guidance
An accessibility concept focused on helping users read dense material with clearer structure, calmer pacing, and less cognitive friction.
Built with: Next.js · FastAPI · Redis · GPT-4o Vision
Experience cues
Peak validation accuracy reported in the MRI classification project.
B.Tech AI & ML at JIET College of Engineering, graduating in 2027.
Production-minded split between product shell and backend services.
Healthcare, diagnostics, accessibility, and internship tooling in one portfolio.
AI Engineering Internships
I am looking for internships where I can ship AI features, build inference pipelines, and work close to real users and product decisions.
Applied ML + Product Roles
The best fit is work that cares about both model quality and product clarity, especially in healthcare, accessibility, or decision-support tools.
Research Collaborations
I am also open to research collaborations where experimentation, validation, and interface design all shape the final result.
Working principles
Step 01
Frame the decision, not just the feature
I start by defining who needs the output, what decision it changes, and where the current process breaks down.
Step 02
Validate the signal before scaling the product
I test whether the dataset, rules, and model behavior actually support the product before I polish the shell around it.
Step 03
Build the product around the intelligence
The model is only one layer. I design the interfaces, state flow, and backend so the intelligence feels usable and easy to check.
Step 04
Measure what works after launch
I evaluate outcomes in terms of clarity, speed, model quality, and whether the product makes the next action easier.
I am an AI/ML engineer focused on turning models into products people can actually use. I care about performance, clear interfaces, and showing enough evidence for others to trust the result.
My work spans clinical tools, accessibility, and automation, with a strong focus on building products that feel practical instead of theoretical.
Kamyavardhan Dave
AI Engineer
Open Console
Direct collaboration channel.
I am actively looking for internships, contract roles, and collaborative projects. Use the console below to understand my current focus, or reach out directly via the channels provided.
Console State / Parameters
Current Phase
3rd year AI/ML student building real products, shipping experiments, and learning through deployment
Current Interests
AI engineering, full-stack AI products, accessibility, automation, and interpretable ML
Response Channel
Email first, LinkedIn for follow-up, usually within 24 to 48 hours
Exploration Mode
Learning by building across domains instead of limiting myself to a single niche too early
Direct Contact
Direct contact, with clear next steps.
Best way to reach me
Email me with the role, project, or collaboration idea and I will reply directly. It is the fastest path and it keeps the site simple and dependable.
Inbox
theway.kamyavardhan@gmail.com
Best for
Internships, product collaboration, applied AI builds
Useful subject line
Role / project / expected timeline
What helps me reply faster
Tell me what you are building, what role you want me for, and whether you need product thinking, implementation, or AI feature design.
Short, clear emails beat long generic intros every time.