Computer vision
AI Seed Detection
Real-time seed identification and classification with a YOLO detector, served through a Python API backed by PostgreSQL.
- YOLO
- CNN
- Python
- PostgreSQL
Aziz KhanForward Deployed Engineer · AI & Full StackHyderabad, India
Aziz Khan, FORWARD-DEPLOYED ENGINEER · AI & FULL STACK: I ship AI systemswhere the customeractually is.
Forward deployed by default: I embed with the team that has the problem, turn ambiguous requirements into working software in days, and harden it into production. 6+ years across Go, Node.js, Python and React; currently building agents, RAG and MCP integrations at Bacancy.
01About
Forward deployed. Full stack. AI-native.
I’m Aziz. For 6+ years I’ve been the engineer you put in the room with the customer: I sit with the people who own the problem, map their systems and data, and ship the first working version fast, then make it production-grade. APIs in Go and Node.js, pipelines and services in Python, interfaces in React.
Today most of my work is applied AI. I design and ship agentic workflows, RAG pipelines and MCP integrations, and I run my own development loop on Claude Code and custom agents so the boilerplate is automated and my time goes to architecture, review and quality.
I care about clear interfaces, boring infrastructure, and code the customer’s team can own after I leave.
0+
Years in production
0+
Projects shipped
0
Companies
Daily stack
GoNode.jsPythonTypeScriptReactFastAPIPostgreSQLAWSDocker
How I work when deployed
From first call to handoff.
Comfortable on-site or remote, in front of engineers and executives, owning the outcome end to end.
- 01→Discover
Sit with the customer, map their workflow, systems and data. Find the real constraint.
- 02→Prototype
Working demo on their data in days, not decks. Agents, RAG, tool-calling, whatever fits.
- 03→Integrate
Wire it into what they already run: APIs, databases, legacy services, auth, MCP servers.
- 04→Harden
Evals, observability, tests, CI/CD. Turn the prototype into something on-call can trust.
- 05Enable
Docs, training, handoff. Their team owns it; I stay on the loop for what comes next.
02Selected work
Things I've built.
A mix of applied ML, backend services and product work.
Edge ML
Gesture Recognition
Hand-gesture model trained from camera input and quantized for on-device inference on ARM hardware.
- Deep Learning
- MMDeploy
- ARM
- PTQ
NLP API
Spell & Grammar Service
Low-latency spelling and grammar correction exposed as a REST service for integration into editors.
- Python
- FastAPI
- REST
Backend service
Email Validation API
Validation pipeline combining syntax checks, MX verification, disposable-domain detection and AI-assisted scoring.
- Python
- FastAPI
- AI
Open source
WishNex
Community platform for sharing wishes and stories, built as a React front end on a Laravel API.
- React
- TypeScript
- Laravel
- MySQL
Browser extension
Visit Tracker
Chrome extension that records browsing time per site and surfaces habits with a small React dashboard.
- TypeScript
- React
- Chrome APIs
WooCommerce plugin
Auto Address Filler
Checkout plugin that auto-completes address fields to cut form friction and abandoned carts.
- PHP
- WordPress
- JavaScript
AI AgentsGoPythonTypeScriptReactFastAPIMCPRAGAWSPostgreSQLAI AgentsGoPythonTypeScriptReactFastAPIMCPRAGAWSPostgreSQL
03Skills
AI-powered speed. Human-grade judgment.
AI-native engineerAgents do the heavy lifting.
Agents do the heavy lifting.
I own the architecture, review and quality.
I build AI agents and applied-AI features for production, and I run my own day-to-day on the same tooling: Claude Code, Cursor and custom agents handle the boilerplate, I make the decisions that matter.
❯
HumanSpec→
AgentsBuild→
HumanReview→
AgentsTest gate→
HumanShip
Claude CodeCursorAI AgentsMCP ServersLLM Tool-CallingRAGPrompt EngineeringAgentic Workflows
Forward deployed
FDEThe engineer who goes to where the problem is. Ambiguity in, shipped system out, customer team enabled.
- Customer discovery & scoping
- Rapid prototyping on real data
- Legacy & API integration
- LLM evals & guardrails
- Demos to execs & engineers
- On-site / remote, any timezone
Discover → Prototype → Integrate → Harden → Enable
Languages
08PythonGoTypeScriptJavaScriptSQLGraphQLBashC/C++
Frameworks
07ReactFastAPIExpressNestJSLaravelFlaskHono
Cloud & DevOps
05AWSDockerLinuxGitCI/CD
ML & Vision
04PyTorchTensorFlowOpenCVYOLO
Working style
05ArchitectureCode reviewMentoringDocumentationOwnership
04Experience
Where I've worked.
— Present
Bacancy
Senior Software Developer · Hyderabad, IN
- Building AI agents and agentic workflows: LLM tool-calling, RAG pipelines, MCP integrations
- Shipping on an AI-native stack (Claude Code, Cursor, custom agents) with tests gating every merge
- Leading full-stack delivery in Go, Node.js and Python from architecture to deployment
- Mentoring junior engineers and owning CI/CD pipelines
—
Logicwind
Senior Software Developer · Surat, IN
- Integrated applied-AI features (LLM APIs, smart automation) into production products
- Designed REST and GraphQL APIs in Go and Node.js
- Cut server response times by 35% through query and async-processing optimisation
—
Logicwind
Software Developer · Surat, IN
- Reduced load times and memory footprint by 25%
- Improved system efficiency by 10% for a 1,000+ user base
—
Logicwind
Laravel Developer · Surat, IN
- Built scalable MVC applications in Laravel
- Optimised data exchange and codebase performance
—
Blockbusters
PHP Developer & Designer · Bhubaneswar, IN
- Cut page load times by 40%
- Maintained 99.9% uptime across client sites
—
BestChoiceHoliday
Web Developer & Designer · Bhubaneswar, IN
- Lifted CTR 10% via A/B testing
- Increased engagement 28% with motion design
- Cut load times by 54%
—
ScanOrder
Full Stack Developer · Bhubaneswar, IN
- Improved order accuracy 15% with analytics
- Reduced deployment time 15% and raised deploy frequency 30%
05Education
Background.
—
Bhubaneswar Institute of Technology
B.Tech, Computer Science & Engineering
Odisha, India
—
Ganapati Institute of Engineering & Technology
Diploma, Electrical Engineering
Odisha, India
06Contact
Let’s build
something together.
Open to Forward Deployed Engineer and senior AI / full-stack roles. On-site, hybrid or remote; happy to travel to where the customer is.
Questions recruiters ask
What does a Forward Deployed Engineer do?
Embeds with a customer or internal team, understands their workflow and data first-hand, ships a working solution fast, integrates it with their existing systems, and hardens it for production before handing it over.
What kind of work are you looking for?
Open to Forward Deployed Engineer and senior AI / full-stack roles. On-site, hybrid or remote; happy to travel to where the customer is.
Which AI systems have you built?
Production AI agents and agentic workflows with LLM tool-calling, RAG pipelines over customer data, MCP server integrations, plus applied ML: YOLO computer vision, on-device gesture recognition, NLP services.
What is your core stack?
Go, Node.js, Python, TypeScript, React, FastAPI, PostgreSQL, AWS, Docker. Comfortable across Python, Go, TypeScript, JavaScript, SQL, GraphQL, Bash, C/C++.
Where are you based and can you travel?
Hyderabad, India. Available on-site, hybrid or remote, across timezones, and happy to travel to where the customer is.