HARDIK
Software engineer building the infrastructure behind automation, the systems behind intelligence, and the software that connects them.

UNSW
Software
Engineering
(Honours)

WAM 82.421. Optiver Industry Prize winner. Letter of recommendation from Director of Studies.
Portfolio showreel thumbnail
PLAY

i'm not just
a developer.

I don't just write code. I architect software. I automate infrastructure. I build AI systems.

Ideas Are Nothing*

without execution

Try to hover the hand
Most people stop at concepts. I build the systems, ship the software, and iterate until it works.
3rd Year @ UNSW Bachelor of Engineering (Software) (Honours)

A quick overview of My Projects

REACT · NODE · POSTGRES · MCP
RECOMMENDED BY SENG DIRECTOR
CoolCoolCool, a B2B SaaS that automates procurement between wholesale and retail. Received a Letter of Recommendation from the Director of studies for software engineering at UNSW. Exposed all endpoints to LLMs through an MCP server as a learning opportunity.
CoolCoolCool screenshot
Live Demo GitHub

B2B E-PROCUREMENT SAAS

AWS λ · S3 · API GW · PYTHON · DOCKER
OPTIVER INDUSTRY PRIZE WINNER
Bravo Constellation, maps 60+ years of market data against complex news sentiment signals. Awarded the Optiver industry prize for elite system architecture and scalability. AWS infrastructure, self learnt within 10 weeks from scratch, simultaneously building.
Bravo Constellation screenshot
Live Demo GitHub

CLOUD FINANCIAL SENTIMENT VISUALISER

FLUTTER · C++ · INFOTAINMENT
Joined the UNSW Sunswift Racing infotainment team to build the next-generation driver interface. Combining solar, battery, and hydrogen fuel cells to push the boundaries of sustainable, clean transport.
Sunswift UI screenshot 1 Sunswift UI screenshot 2 Sunswift UI screenshot 3 Sunswift UI screenshot 4

Sunswift Infotainment UI

Vercel Serverless Architecture
SENTIMENT PROBABILITY MODEL (SPM)
Want the full picture?

Flip for
the CV.

Heads - you hire me. Tails - you hire me anyway.
The coin always lands in your favour.
Full-Stack · AI/ML · Cloud

The Stack

Tools I ship with the best.
Find more in my CV
Coverage
0.93
SYSTEM DESIGN
Distributed Systems
System Architecture
REST APIs
Design Patterns
CLOUD & INFRASTRUCTURE
AWS
Docker
CI/CD
Linux
AI & AUTOMATION
MCP
Reverse Engineering with AI
LLM APIs
Agentic Workflows
SOFTWARE ENGINEERING
Python
TypeScript / JavaScript
PostgreSQL
Java
UNSW Software Engineering (Honours) · WAM 82.421

FULL-STACK · AI/ML · CLOUD

sustainability

Not new to shipping. From React
frontends to AWS microservices to LLM
pipelines with real tool calling I build
things end-to-end.

END-TO-END
OWNERSHIP

I don't give up when it gets hard. From scoping to shipping to post-deploy - I stay on the problem until it's actually solved.
1

Upskilling is
not a blocker

New stack, new domain, new constraints - I've taught myself everything from AWS to creating a MCP server mid-project. Give me a week and I'll have opinions.

FAST.
NOT SLOPPY.

Speed and quality aren't a trade-off if you build the right habits early. CI, test coverage, clean PRs - I treat them as defaults, not afterthoughts.

I learn from failure,
and improve with every iteration.

Here's a snapshot of examples of what I've built and learnt alongside.
Lesson
Project
"Picked AWS Lambda for BRAVO because it felt like the easiest choice. Then spent six weeks learning that the infrastructure is never the easy part
The code was done in two weeks.
The deploy took the rest."
Bravo Constellation, Financial data API (2026)
"I started writing tests because the spec required 80% coverage.
By the final sprint I was writing code that existed to make the tests easier to write. 93% coverage is the reason we could ship without breaking code."
CoolCoolCool, AI B2B SaaS (2026)
"Building the MCP server taught me that the consumer redesigns the contract When a LLM is calling your tools, ambiguous parameters cause failure. Designing for AI consumption has zero tolerance for vagueness."
CoolCoolCool, AI B2B SaaS (2026)
"The credit card sequence planner is greedy. Clean, fast, explainable. But greedy misses dynamic selection cases. Knowing when your algorithm is good enough is an engineering decision."
OpenLoans, Credit Card Signup-Bonus Manager (2026)
"At Sunswift, a software decision that wastes 0.3% battery capacity costs 10 miutes in a race you spent a year building for. I've never thought harder about the cost of an abstraction. That is the fastest way to care about your code."
Sunswift Racing, UNSW solar car (2026)

Pick a project

Price x News AI B2B AaaS OpenLoans

Financial sentiment & market event visualiser. Three-microservice architecture on AWS, star-map UI for sentiment vs price where each trading day is a star in space. Won the Optiver industry prize at UNSW 3011 Capstone, 2026.
Full-stack production storefront natively integrated with a custom Model Context Protocol (MCP) server. Exposes 29 secure REST endpoints as machine-readable tools, allowing LLM agents to execute end-to-end user workflows using natural language. Earned a letter of recommendation from the UNSW SENG Director of Studies.
Financial optimization simulator and "Points Hacking" assistant designed and shipped from scratch in under 48 hours for a rapid engineering sprint. Features a low-level GPGPU physics layer, shifting thousands of compounding vector math calculations directly onto GPU shaders.
Python · AWS Lambda · API GW · S3
React · Node.js · PostgreSQL · MCP
Next.js 14 · WebGL/GLSL · Zustand
Microservice cloud architecture
29 REST endpoints, 93% coverage
GPGPU particle physics rendering
Won Optiver industry prize
Director of Studies recommendation
48-Hour high-velocity rapid prototype
2026

Sentiment Visualiser
2026

E-Commerce Platform
2026

OpenLoans
Three microservices, one star map. Sentiment in, signal out. A storefront engineered like real product infra. Algorithmic strategy simulation meets zero-latency browser physics.
Stack:
Python · AWS · Docker
Stack:
React · Node · PostgreSQL
Stack:
Next.js · WEBGL · Zustand
Surface:
Interactive web visualiser
Surface:
Full e-commerce frontend
Surface:
GPGPU Particle Simulator
Backend:
3 microservices on AWS
Backend:
29 REST endpoints
Backend:
Deterministic Math Engine
Tests:
Unit + integration
Tests:
93% coverage
Tests:
User Acceptance Loop
Recognition:
Optiver Industry Prize
Recognition:
Director of Studies letter
Recognition:
Shipped Under Strict deadline
Year:
Feb–Apr 2026
Year:
Mar–Apr 2026
Year:
May-May 2026
Repo:
github.com/Hardik-V/3011_Bravo
Repo:
github.com/nloudev/H14B_CoolCoolCool
Repo:
github.com/Hardik-V/OpenAptitude
Abstract
Hardik, a UNSW Software Engineering (Honours) student specialising in AI agents, cloud microservices, and full-stack systems. Trained on Python, TypeScript, Java, C, and SQL. Fine-tuned on AWS Lambda, API Gateway, S3, PostgreSQL, React, and a healthy obsession with Docker. Released open-source with reproducible projects, working repos, and a paper trail of recommendations. Limitations include occasional over-engineering and an inability to leave a failing CI pipeline alone.
Education
Bachelor of Engineering (Software) (Honours) · UNSW Sydney · 2024–2027 · WAM 82.421. Relevant coursework: Artificial Intelligence, Algorithms & Data Structures, Software Testing & QA, Extended Security Engineering, Database Systems, Computer Networks, Algorithm Design & Analysis.
Recognition
Winner of the annual Optiver industry prize for excellence at UNSW SENG3011 Capstone, 2026. Received a letter of recommendation from Fethi Rabhi, Director of Studies for undergraduate Software Engineering at UNSW, for exceptional project quality on the E-Commerce & AI Automation platform.
Skills
Languages:   Python, Java, JavaScript, TypeScript, C, SQL
AI & LLM:    Claude, OpenAI, MCP, Agentic Workflows, API
Backend:     Socket, Express, REST APIs, PostgreSQL, Jest
Cloud:       AWS (Lambda, API Gateway, S3, EC2), Docker
Frontend:    React, HTML, CSS, WebGL/GLSL, Three.js
Concepts:    Architecture, OOP, DSA, Networks, Databases
Hardik · Verma