Ankush Kamble
Available for opportunities

Ankush Kamble

Engineering Resilient Backend Systems & AI Products

Backend-focused Software Engineer based in Toronto, CA. 4+ years building Python APIs, production infrastructure, and autonomous agent workflows.

Open to Full Stack · Backend · Frontend roles
Follow
Scroll

where i've worked

The Core Track Record

Tap any card to reveal the impact metric

Software Engineer - Core Infrastructure & Tooling

Aquent Canada

Aug 2021 - Present

Full-TimeToronto, ON

CI/CD Pipeline Engineering

Shipped multi-environment CI/CD pipelines using GitHub Actions and Jenkins to eliminate release bottlenecks and accelerate deployment velocity.

GitHub ActionsJenkinsCI/CD

Python Microservices Architecture

Designed Python/REST API microservices to synchronize automated runtime diagnostics with project management APIs, enabling real-time observability.

PythonREST APIsMicroservices

Production Reliability & Hotfixes

Optimized production backend performance, executing high-priority hotfixes to protect user retention under active production conditions.

HotfixesPerformanceReliability

things i've built

What I've Shipped

Two things I built from scratch - an AI skincare companion and a resume agent that actually works.

Featured Project

Nura

AI Skincare Companion

Founder & Lead Engineer
Pick skin type

Select your skin type

DAU/MAUTracked
4 × 4Skin matrix
OpenAIVector pipeline

Integrated OpenAI API pipelines for context-aware vector embeddings; tracked DAU/MAU retention analytics to guide backend query tuning for performance-sensitive AI flows.

SwiftUIFastAPIPythonSQLOpenAI APIVector Embeddings
Featured Project

Resume Tailor

Autonomous Agent & Web Interface

Full-Stack Engineer

Drop your resume or click to simulate

PDF · DOCX · TXT

↑ Drop or click to trigger the agentic self-correction pipeline

3+Slop flags/run
CLI→APIRefactored
PydanticSchema-safe

Converted a Python CLI pipeline into a decoupled FastAPI REST service with Pydantic schemas and a type-safe React frontend. Autonomous quality gates detect 3+ anti-slop failures per run, run targeted LLM refinement passes, and validate evidence density.

ReactTypeScriptViteTailwind CSSFastAPIPydanticLLMs

tools i reach for

The Full Stack

Tools and technologies I reach for when building things that have to actually work.

Underlined skills link to where I've applied them

Languages & Frameworks

95%
90%
Django80%
75%
82%
78%
SQL85%

Cloud & Infrastructure

90%
Postgres88%
Supabase85%
Firebase80%
Stripe72%
95%

DevOps & Tooling

90%
88%
82%
Selenium78%
Docker75%
85%

how it works

System Design

Resume Tailor agent loop - from raw input to quality-verified output.

Run the pipeline or tap any node to inspect its role

fail ↓

↑ retry

Press Run Pipeline or tap a node to explore

Input layer
Processing
LLM + Output
Self-correction loop

say hello

Don't be a Stranger

Hiring, building something, or just want to talk backend - my inbox is open.

ankushkamble1999@gmail.com

Schedule a 30-min call

Calendly · Free slot