Loading Experience...
AI-FIRST SOFTWARE ENGINEER
I'm Aryan Kashyap, a Software Developer & ML Engineer
building production-grade APIs, automation, and analytics around intelligent models.
I sit at the intersection of ML, backend engineering, and cloud infra
re‑architecting retail systems at Aptos and shipping data-heavy SaaS at EasyInsights.
Curious about the story behind the work? Explore my experience, stack, and impact.
Currently · Software Developer L1 @ Aptos Retail
Elsewhere
I’m currently a Software Developer L1 at Aptos Retail, re‑architecting high‑traffic order and product APIs, modernising legacy services, and squeezing cost and performance out of a large AWS microservices estate. Before that, I worked as a Full Stack Developer at EasyInsights on data‑heavy SaaS, and as a QA Automation Engineer at Schlumberger building robust test automation.
My work sits at the intersection of ML, backend engineering, and cloud infrastructure - from ML‑driven products to API design, observability, and regression frameworks that keep systems safe to change. I studied Computer Science with specialization in AIML at VIT Bhopal (CGPA 8.56), where I focused on applied AI.
Prefer email? akmail2017@gmail.com
A few systems and experiments that show how I think about ML products, infra, and orchestration.
I've watched people struggle with scattered medical records across dozens of systems. CareLens solves that, transforming PDFs, wearables data, and lab reports into one intelligent timeline. The platform normalizes data from EHR systems and manual uploads, uses Gemini for trend analysis and anomaly detection, and gives patients a complete view of their health journey. Competed solo and landed Top 10 among 300+ teams (max 2 per team).
🔒 Enterprise
An AI copilot that actually understands your infrastructure. Built a multi-modal orchestrator using Llama 4 that pulls context from Jira, Confluence, and Bitbucket via hybrid RAG with Qdrant. It doesn't just answer questions. It creates tickets, deploys infrastructure, and diagnoses faults autonomously. Cut mean time to resolution by 67% across services.
Production translation pipeline that handles the messy reality of enterprise documents: XML, JSON, CSV, PDF, you name it. Hybrid approach combining LLMs with Amazon Translate, smart batching for 65K+ strings, and back-translation QA to catch semantic drift. Hitting 92% accuracy with an 87% similarity threshold that actually preserves meaning.
Precision agriculture for farmers who need it most. Computer vision models hitting 98.2% yield prediction and 99.3% disease detection. The real challenge? Making it work offline in rural areas with spotty connectivity. Built a mobile-first dashboard with multilingual support so farmers can actually use it in the field.
We've all been there. You know the command exists, you just can't remember the exact syntax. Built this as a searchable reference for 54+ languages and tools. No more digging through Stack Overflow for that one Git command you use twice a year.
A collection of wild frontend experiments I built while learning. No frameworks, no dependencies, just vanilla HTML, CSS, and JavaScript pushing the limits of what's possible in the browser. Every project taught me something new about the web platform.
From prototype notebooks to production APIs, I care about fast feedback loops, observability, and deployments that stay boring in production.
Python
TypeScript
TensorFlow
Pandas
Hugging Face
LangChain
FastAPI
React
AWS
Docker
Kubernetes
PostgreSQL
Terraform
Grafana
Prometheus
Redis
Git
Tableau
Real-time visualization of my coding activity. Watch the commits flow.
@aryankashyap7
Arcade · dodge the debris
Pilot your spacecraft through an endless asteroid field. How long can you survive the chaos of deep space?
Press any arrow key to start
