Open to applied AI roles & consulting

Ashwin
Balasubramanian

// Applied AI Engineer

I build practical AI products across model behavior, backend systems, and infrastructure — prompt engineering, agent workflows, and production reliability that improve quality, reduce cost, and solve real user problems.

01/About

About Me

I build applied AI systems that are useful in production, reliable under pressure, and grounded in real product outcomes.

At Vizly, I work across prompt design, agent behavior, backend systems, and cloud infrastructure to improve how AI products perform in the real world. My recent work includes building secure MCP integrations with OAuth, shipping enterprise prompt-governance features, and reducing infrastructure costs without sacrificing reliability.

I enjoy working at the layer where model quality, product experience, and engineering rigor meet. Whether I’m debugging production issues, designing data workflows, or building AI-powered developer tools, I care about making systems practical, measurable, and easy for people to trust.

Southfield, MI (Remote)

Snapshot

Current Role

Applied AI Engineer at Vizly.ai, Inc.

June 2025 – Present

M.S. Computer Science

Arizona State University

Aug 2023 – May 2025 · Tempe, AZ

B.Tech Computer Science & Engineering

Amrita Vishwa Vidyapeetham

Jul 2017 – Jun 2021 · Coimbatore, India

02/Skills

Skills & Technologies

A comprehensive toolkit for building intelligent data-driven solutions

Programming Languages

PythonJavaScriptSQL (Postgres)HTML/CSSJavaR

AI & LLM

Prompt EngineeringOpenAIAnthropic ClaudeGoogle GeminixAI GrokLangChainVercel AI SDKRAG

Frameworks & Backend

FastAPINext.jsReactFlaskNode.jsTensorFlowPyTorch

Cloud & DevOps

AWSTerraformKubernetesDockerCI/CDGitECSEKS

Data & ML

SupabasePandasNumPySciPyScikit-learnNLTKSpaCyXGBoost

Platforms & Tools

n8nLangfuseStripe APIVS CodeAPI DevelopmentModel Deployment

03/Projects

Featured Projects

Real-world projects spanning AI, data science, and software engineering — solving complex business and technical challenges

01

MacroSense AI

In Progress

Feb 2026 – Present

Architecting a multi-tenant SaaS health platform with provider-agnostic AI routing, enterprise RBAC, Postgres RLS, and Langfuse observability. Engineering Fitbit OAuth 2.0 PKCE integration with encrypted token storage and resumable 90-day Inngest backfill workflows, alongside cache-first food search and privacy-focused data export/deletion flows.

Key Features

Multi-Tenant SaaS Architecture
Fitbit OAuth 2.0 PKCE
Langfuse Observability
GDPR-Compliant Data Workflows

Technologies

Next.jsReactSupabaseInngestVercel AI SDKAnthropicOpenAILangfuseFitbit API
02

BioCAD (Biomedical Computer-Aided Diagnosis)

Research

Aug 2023 – Dec 2023

Trained a unified deep learning pipeline across CheXpert, ChestX-ray14, MIMIC, VinDr-CXR, and ChestX-Det to classify, localize, and segment chest X-rays using a student-teacher contrastive learning architecture with cyclic pretraining to mitigate catastrophic forgetting. Applied ConvNeXt for classification, Faster R-CNN-ResNet for localization, and UNet-Swin for segmentation, achieving AUC 0.7348 on ChestX-ray14, mAP 0.115 on VinDr-CXR, and Dice 0.4711 on ChestX-Det using 2x A100 CUDA GPUs. Developed as part of a joint medical imaging research initiative between Arizona State University and Mayo Clinic, contributing deep learning solutions to assist radiologists in early and accurate detection of pulmonary conditions.

Key Features

Student-Teacher Contrastive Learning
Cyclic Pretraining Across 5 Datasets
AUC 0.7348 on ChestX-ray14
mAP 0.115 and Dice 0.4711

Technologies

PyTorchTensorFlowPythonCUDAComputer VisionContrastive Learning

04/Experience

Professional Experience

Building intelligent systems and leading data-driven initiatives across various industries

Applied AI Engineer

Vizly.ai, Inc.

June 2025 – Present

Southfield, MI (Hybrid)

  • Improved agent output quality by tuning prompts and model behavior for stronger user-intent understanding, cleaner code generation, and fewer customer-reported quality issues.
  • Built a production Model Context Protocol server with OAuth 2.0 so AI agents could securely invoke data integrations across the platform.
  • Re-architected JupyterHub infrastructure with Terraform and AWS, reducing monthly cloud spend by 40% while maintaining 257 ms execution latency and zero downtime incidents.
  • Developed enterprise prompt-governance capabilities for teamspaces, including inherited custom prompts, conflict resolution, and guardrail-based policy alignment.
  • Resolved 12+ production and P0 issues spanning billing, authentication, and agent response behavior to improve platform reliability and customer experience.
OpenAIAnthropic ClaudeAWSTerraformOAuth 2.0MCPPrompt Engineering

Data Science Summer Intern

Aggregate Intelligence Inc.

May 2024 – Aug 2024

Remote

  • Built an AI assistant using OpenAI APIs and company REST APIs to deliver natural-language responses for customer interactions.
  • Developed a Next.js-based chatbot interface that improved usability and streamlined end-user interactions.
PythonOpenAI APINext.jsReactREST APIs

Data Analyst and Python Developer

Aggregate Intelligence Inc.

Sep 2022 – May 2023

Coimbatore, India

  • Built data-driven RV park market scoring models using statistical and geospatial methods to support strategic business decisions.
  • Integrated demographic, economic, and geographic datasets to improve signal quality and derive actionable market health insights.
PythonStatistical AnalysisGeospatial AnalysisData IntegrationMarket Scoring

Data Science Intern

Algolitics India LLP

May 2019 – Jun 2019

Coimbatore, India

  • Processed high-volume self-storage data and performed longitudinal trend analysis over a 7-year period.
  • Produced analytical reports with pricing and promotion insights for cross-functional product stakeholders.
PythonData AnalysisTrend AnalysisReporting

05/Contact

Get In Touch

Interested in applied AI, intelligent product work, or building something useful together? Reach out and let's start the conversation.

Send a Message

Share a bit about what you're building, hiring for, or exploring and I'll take it from there.

0/2000

Complete the verification to protect this form from spam.

Usually responds within 24 hours