Currently building → AYNX / Optivara

Building systems for
better decisions.

I’m Aayuush — an AI & Data Science undergraduate at IIT Jodhpur, founder of AYNX, and builder exploring the intersection of machine learning, mathematical optimization, decision intelligence, and quantum computing.

Explore the workLet’s connect
LinkedIn GitHub Email
DATAPATTERNSCONSTRAINTSOPTIMIZATIONDECISIONIMPACT

01 — About

I don’t want to just predict the future. I want to optimize what happens next.

I study AI & Data Science at IIT Jodhpur, and I approach technology mostly by building and experimenting with it. What holds my attention is a specific gap: the distance between “we have data” and “we made the right decision”.

Machine learning helps us understand what might happen. Optimization asks a harder question — given what we know, what should we actually do? Objectives, constraints and uncertainty all live in that second question, and most real systems fail there rather than at the model.

That’s the thread connecting everything here: AYNX, Optivara, the open-source work, and the quantum exploration are chapters of the same problem, approached from different directions.

The intellectual map

BETTER DECISIONS

Hover a node to see what it means, why it matters here, and where it shows up in the work.

02 — Building

AYNX

Decision intelligence for complex systems.

Organizations don’t always need more data. They need better ways to transform data, constraints, uncertainty and objectives into decisions. AYNX is my attempt to build that layer deliberately, starting with the problems where the constraints are the hard part.

Optivara

Decision optimization engine

Optivara is a decision optimization platform aimed at logistics and operations: routing, scheduling, resource allocation and operational planning — the constraint-heavy problems where a good forecast still leaves you without a plan.

  • Python
  • OR-Tools
  • Constraint programming
  • Mathematical optimization
  • Predictive analytics
  • Systems modeling

Status: In development — prototype stage

Locations, demand and vehicles — unstructured.

03 — The next frontier

Classical → Quantum

QForge is my exploration track — not a product, not a claim of expertise. It sits at the point where optimization, quantum computing and quantum-inspired methods meet, and it exists because of one question:

What happens when the computational paradigm itself changes?

Classical solvers eventually hit structural limits on combinatorial problems. Whether a different substrate meaningfully moves those limits is an open empirical question — one I’m studying, not answering. No quantum advantage is claimed here.

Status: Exploring — foundations & prototypes

  1. 01Classical optimization
  2. 02Heuristics
  3. 03Quantum-inspired optimization
  4. 04Quantum algorithms
  5. 05Quantum-AI

04 — Proof of work

Ideas are cheap. Repositories are evidence.

Selected work, written up as small case studies rather than cards. Everything below lives publicly on github.com/Aayuush1.

CinemaMatch

  • Python
  • NumPy
  • Linear algebra
  • SVD
  • Recommender systems
View GitHub
The problem
How do you recommend something a person hasn't seen, using only the sparse traces of what everyone else has?
The approach
A recommendation system built on singular value decomposition — factorizing the user–item matrix into latent structure instead of hand-written rules.
The learning
Latent factors are only as honest as the sparsity underneath them. Evaluation design matters more than model choice.

Customer Support OpenEnv

  • Python
  • LLM agents
  • Environment design
  • Evaluation
View GitHub
The problem
Agentic systems are usually demoed, rarely measured. What does a controlled environment for them look like?
The approach
An environment-oriented AI project: framing customer support as a structured setting an agent acts within, so behaviour can be observed rather than asserted.
The learning
Most of the difficulty in agents is not reasoning — it's specifying the boundary between what the system decides and what a human still owns.

Quantum Foundation

  • Python
  • Quantum computing
  • Linear algebra
View GitHub
The problem
Before asking whether quantum helps optimization, you have to actually understand the primitives.
The approach
A study repository working through quantum computing fundamentals — states, gates, circuits — from first principles rather than analogy.
The learning
The intuition gap is mathematical, not mystical. Progress here is slow, and that is the honest reporting.

Capabilities

Computation

  • Python
  • SQL
  • Git
  • Jupyter

Data

  • NumPy
  • Pandas
  • Scikit-learn
  • Data visualization

Intelligence

  • Machine learning
  • Data science
  • Generative AI

Optimization

  • OR-Tools
  • Constraint programming
  • Mathematical optimization

Frontier

  • Quantum computing
  • Quantum-AI

05 — Journey

The trajectory, so far.

  1. AYNX

    Founder

    Building decision intelligence systems for complex, data-driven environments.

    2026 — Present

  2. SkillPraxis

    Founder

    A hands-on learning initiative centered on capstone projects and real-world problem solving.

    2025 — Present

  3. Open Source Connect

    Contributor

    2026

  4. GirlScript Summer of Code

    Contributor

    2026

  5. Scaler School of Technology

    Ex-Intern

    2025

  6. IIT Jodhpur

    Undergraduate — AI & Data Science

    2025 — Present

Education

Indian Institute of Technology Jodhpur

B.S. Artificial Intelligence & Data Science · 2025 — Present

Foundation in computational thinking, mathematics, statistics, Python, SQL, data science, machine learning and analytical problem solving.

Timeline

  1. 2021

    PCM foundation

  2. 2025

    IIT Jodhpur

  3. 2025

    Scaler School of Technology

  4. 2025

    SkillPraxis

  5. 2026

    AYNX

  6. 2026

    Optivara

  7. Now

    AI + Optimization + Quantum exploration

  8. Next

    Intentionally unwritten.

Selected recognition & programs

  • India Genius Award — Level 1
  • Claude Code in Action
  • Snowflake Data for Breakfast
  • India AI Impact Buildathon
  • Gen-AI Mastermind
  • GSSoC'26
  • OSCG'26
  • OSCI'26

06 — Thinking

Thinking in public.

An evolving notebook — writing on AI, agents, optimization, quantum computing and what building AYNX actually teaches. Published on LinkedIn as it’s written.

Article · LinkedIn

The Agentic Trap

On the gap between autonomous-looking systems and systems that actually make reliable decisions — and why agency without constraints is not the same as capability.

Read on LinkedIn

$ whoami

Aayuush — building at the intersection of AI, optimization & computation.

Writing tracks

  • AI & agents
  • Optimization
  • Quantum computing
  • Building AYNX
  • Technical philosophy

Currently exploring

status: exploring

Decision Intelligence

How can AI move beyond prediction toward action?

Optimization

How can mathematical models solve difficult real-world resource allocation problems?

Quantum Computing

Can emerging computational paradigms change the limits of optimization?

AI Agents

Where should autonomous systems act — and where should humans remain in control?

Depth over
shortcuts.

  • First principles over assumptions.
  • Build over consume.
  • Systems over isolated features.
  • Substance over hype.
  • Long-term thinking over short-term validation.
  • Theory should eventually meet reality.

07 — Contact

Have a hard problem?

I’m interested in ambitious technical problems, AI, optimization, quantum computing, open source, research, and people building for the long term.

Email Aayuush LinkedIn GitHub

Aayuush@proton.me