01
AI / RAG
Multimodal Financial RAG
Bloomberg data pipeline with BLIP-2 image + text embeddings for intelligent multi-asset portfolio analysis.
Python · BLIP-2 · LLaVA · RAG · Bloomberg API
↗
Available Summer–Fall 2026
Building Intelligent Systems
›
CS junior at Arizona State University. I build scalable backend systems, serverless APIs, and AI tools — from neural networks to trading bots to real-time telemetry for rovers in the desert.
AI / ML
Backend & APIs
Agentic AI
Trading Automation
Cloud & Serverless
GPA
0
Arizona State University
Current Focus
ML pipelines · Agentic systems · GPU-accelerated inference · Scientific ML at Rolston Lab
Based in
Tempe, Arizona
scroll
01 — About
Hi, I'm
Aayush
I'm a Computer Science junior at Arizona State University (GPA 3.84) building systems at the intersection of AI, backend infrastructure, and scientific computing.
I take ideas from experiment to production — with strong observability, clean APIs, and real-world reliability. Currently doing ML research at Rolston Lab and actively seeking SWE & AI/ML internships.
- BLIP-2 pipelines for multimodal scientific & financial data
- GPU-accelerated ML workflows & latency-optimized inference
- Real-time telemetry dashboards for robotics & rovers
- Agentic trading systems integrated with MT5
Coursework
02 — Skills
What I build
with
01
Languages
Python · TypeScript · JavaScript · Java · C/C++ · Go · SQL · Swift · MATLAB
02
AI / ML & Data
PyTorch · TensorFlow · Hugging Face · BLIP-2 · LLaVA · OpenAI APIs · NumPy · Pandas · OpenCV
03
Backend & APIs
FastAPI · MERN · Node.js · RESTful APIs · Serverless · Edge Runtime · GraphQL
04
Infrastructure
PostgreSQL · MongoDB · Firebase · InfluxDB · Cloudflare Workers · AWS · Google Cloud
05
DevOps & Practices
Docker · Kubernetes · GitHub Actions · CI/CD · GPU Clusters · ML Ops · Agile/Scrum
03 — Experience
Where I've
shipped
ML Researcher
Rolston Lab · Arizona State University
- Developing a 4D ML model for interfacial phenomena in energy materials
- Built BLIP-2 VQA pipelines processing multimodal scientific datasets
- Designed GPU-accelerated training workflows with Gaussian Process Regression
Python · PyTorch · BLIP-2 · GPR · GPU Clusters
AI Backend Developer
Startup (Remote)
- Built serverless AI microservices on Cloudflare Workers with sub-50ms cold starts
- Integrated LLM APIs with custom retrieval pipelines for product features
- Designed edge-native architecture with Durable Objects for stateful inference
Cloudflare Workers · TypeScript · LLMs · Edge Runtime
Rover Telemetry Engineer
Sun Devil Robotics · ASU
- Built a real-time telemetry dashboard for the rover competition team
- Integrated InfluxDB + Grafana for live sensor stream visualization
- Implemented Socket.io backend for low-latency rover command relay
Next.js · TypeScript · Socket.io · InfluxDB · Grafana
04 — Projects
Things I've
built
02
Robotics
Rover Telemetry Dashboard
Real-time telemetry & command dashboard with live sensor streams, used in desert rover competitions.
Next.js · TypeScript · Socket.io · Grafana · InfluxDB
↗
03
Scientific ML
Battery Health Prediction
Gaussian Process Regression pipeline using EIS data to forecast lithium-ion battery degradation.
Python · SciKit-Learn · GPR · EIS · Scientific ML
↗
04
Trading Automation
Elite Confluence Trading Bot
Fully automated XAU/USD strategy integrating multi-indicator confluence signals with MT5 execution.
Python · MT5 · Technical Analysis · Automation
↗
05
Voice AI
Cloudflare AI Voice Assistant
Voice-enabled assistant on Workers AI using Llama 3.3 with Durable Objects for stateful low-latency inference.
Cloudflare Workers · Workers AI · Llama 3.3
↗
06
Mobile / iOS
Weather & Cafe Finder
iOS app suggesting nearby cafes and outfit choices using real-time weather + geolocation with Firebase auth.
Swift · iOS · Firebase · REST APIs
↗
05 — Research
Rolston Lab
work
R1
4D ML Model for Interfacial Phenomena
ML models predicting interfacial behavior in energy materials across space, time, and operating conditions — enabling smarter energy system design.
R2
BLIP-2 Scientific Data Extraction
BLIP-2 pipelines extracting structured information from microscopy images, plots, and scientific figures for high-throughput automated preprocessing.
R3
Battery EIS ML Pipeline
ML workflows using Electrochemical Impedance Spectroscopy (EIS) for battery health prediction and degradation modeling with Gaussian Process Regression.
06 — Activities
Core Team Member
Google Developer Group · ASU
Organizing technical workshops on ML and web dev for 100+ students.
Adobe Student Ambassador
Adobe
Beta testing AI features in Adobe Express — UX, usability, and creative workflow feedback.
07 — Contact
Let's build
something real
Open to SWE & AI/ML internships, research collaborations, and product ideas. Reach out.