Full-Stack AI Engineer

Turning intelligenceinto useful systems.

Hammad Asghar. I build systems where models, agents, tools, retrieval, APIs, and data come together to solve a real problem — with a refusal path when they cannot.

Scroll to construct a system

SELECTED WORK

Engineering workwith a point of view.

01 / 04

Public website

Naseebi — Website

naseebi.com presents Naseebi as a marriage and matchmaking site for people and families: verified profiles, private messages and calls, and family involvement.

View case study →
Naseebi sign-up screen asking for an email address, with Get OTP, Continue with Google, and Continue With Phone.
03 / 04

Personal prototype

ReconcileOps

A payment reconciliation prototype: CSV import, deterministic AED matching, exception investigation, and an audit trail. A React client sits on a Nest API.

  • React
  • Vite
  • NestJS
  • PostgreSQL
View case study →
ReconcileOps reconciliation screen with filters for matched, amount mismatch, missing bank entry, missing payment, ambiguous, and outside the settlement window.
04 / 04

Public CRM

Fast Holidays CRM

crm.fastholidays.org is a client-rendered app titled Fast Holidays. Supplied screens from that website show sign-in and an employee portal for attendance and leave. Earlier artwork in the gallery is branded FastTrack.

View case study →
Fast Holidays sign-in screen. The panel reads Welcome back, with email and password fields, and a note about clients, visa orders, and invoicing.

Also on this site

LIVE SYSTEM MODEL

How the pieces connect.

LIVE SYSTEM MODEL01 / 04
01Intentuser input
02Reasoningstructured plan
03Tools / MCPcapability layer
04Contextretrieved knowledge
05Datasource of truth
06ServicesAPI / workflow
07Outcomeproduct action

THE PRACTICE

A model is only onepart of the product.

The hard work lives in everything around it: good inputs, deliberate workflows, dependable data, thoughtful interfaces, and a clear way to understand what happened.

01

Useful by design

Start with a real task, then shape the AI around the people doing it.

02

Inspectable in practice

Make sources, tool calls, and tradeoffs legible—not magical.

03

Built to evolve

Make the first release robust enough to learn from, then improve it with evidence.

AI ENGINEERING LAB

The system, madelegible.

Small, honest demonstrations of the engineering patterns. Interactive demos — not a live model, not invented scores.

GROUNDED ANSWERS

Retrieval before generation.

A walkthrough of the Corpus pattern: query, retrieve, assemble context, then extract or refuse. Interactive demo — not a live model.

  1. 01
    Query analysisIntent: answer quality / guardrails
    ready
  2. 02
    Hybrid retrievalBM25 always; dense only when keyed
    waiting
  3. 03
    Context assemblyRanked, attributed chunks
    waiting
  4. 04
    Answer with citationsExtract or refuse — no fluent guess
    waiting
EXAMPLE CHUNKS

TECHNOLOGY MAP

Every layer has totalk to the next.

SELECT A SYSTEM

Used in case studies, currently learning, or exploring — labeled honestly.

CURRENT TRAJECTORY

Building the edge.

A working roadmap — not a skill bar and not a claimed production stack.

NOW

Python for LLM tooling
Hybrid retrieval
Agents & MCP shape

NEXT

AI evaluation
Observability
AI infrastructure

EXPLORING

Multimodal
Voice agents
Advanced orchestration

LET'S BUILD SOMETHING THAT WORKS

Have a systemworth thinking about?

asghar.hammad110@gmail.com ↗
© 2026 / Independent engineeringGitHub · LinkedIn · Back to top ↑
PORTFOLIO NAVIGATOR
Structured search across systems, projects, and practices
Project
Naseebi — WebsiteMatchmaking · Website
↗
Project
Naseebi — Mobile AppReact Native · Expo · Agora
↗
Project
ReconcileOpsNestJS · PostgreSQL · AED
↗
Project
Fast Holidays CRMCRM · Clients
↗
System
Corpus — retrieval, not guessworkRAG · BM25 · PostgreSQL
↗
Project
Helix — visible tool callingAgents · MCP · Zod
↗
Project
Fieldkit — React Native operationsReact Native · API · Outbox
↗
Practice
Inspectable AI systemsCitations · Refusal · Demos
↗
↑↓ navigate↵ openESC close
ASK MY AI

Same-origin Ask → Next `/api/agent/run` → Nest catalog agent. Lexical Postgres FTS until embeddings exist. Chat completions when the OpenAI key is billed; otherwise local-extractive. Cited or refused. Secrets stay off the client.

Checking catalog API…

⌕
Nest catalog on :4000Cited or refusedESC close

ARCHITECTURE EXPLORER

A system with
clear boundaries.

Choose a node to see why it exists and how it participates in the flow.