AuraX

AI Researcher & CTO·Hyderabad, India·Dec 2024 – Dec 2025

What they do

  • AuraX was a fashion-tech company incubated at IIIT-Hyderabad's CIE. Fashion e-commerce needs on-model photography for every garment, in every size, on every body type, and a photoshoot per SKU does not scale. We generated that imagery with diffusion models instead.

The work

  • I joined as AI researcher and CTO and ran the research. The problem sounds like image generation and is not. A customer already knows what the shirt looks like, so the garment is a constraint rather than a suggestion, and getting it approximately right is the same as getting it wrong.
  • The gap nobody was serving was non-Western clothing. Open models are trained mostly on Western casual wear, so they handle a t-shirt well and fall apart on a saree, because pleats and draping physics are not something you can approximate from a prior built on tubes of fabric. We built for that case specifically, and it turned out to be the most defensible thing we had.
  • Alongside the research I built the production side: backend and GPU inference across AWS, ShaktiCloud and Modal, with latency tuned until brands would actually put it in front of customers. Modal covered the serverless side, where inference is bursty and paying for an idle GPU makes no sense, while the sustained training and R&D work sat on dedicated nodes. Research quality and shipped quality are different problems, and the second one is less fun and takes longer.
  • It worked. In blind tests people picked our images as real over Google's and OpenAI's image models, and on Nautinati, Aditya Birla Fashion's kidswear brand, our images got a 3% click-through rate against 1.5% for studio shoots. We closed deals with Mensa Brands, Aditya Birla Fashion and TTK Healthcare. Real contracts, real revenue.
  • Then the next round didn't come in time, and the big labs moved into the space. We were offered ₹28.5 lakh more. It would have kept the company open, but it couldn't buy the compute to keep our quality lead, and staying open without that lead wasn't worth it. We said no and closed it.
  • The lesson I actually took is not about diffusion models: being right about the technology and being right about the business are separate bets, and I had been treating a win in the first as evidence about the second. Distribution was the bet we never placed, which is why it's the whole point of what I'm building now.

What I did

  • Joined as AI researcher and CTO in Dec 2024. AuraX was incubated at IIIT Hyderabad's CIE and raised ₹22 lakh from CIE, IIM Bangalore's NSRCEL and MeitY's TIDE 2.0.
  • Built an aesthetic fine-tuning loop for fashion image generation: generate, score for aesthetics and garment accuracy, feed the scores back, repeat. In blind tests people picked our images as real over Google's and OpenAI's image models.
  • Closed deals with Mensa Brands, Aditya Birla Fashion Ventures, Aditya Birla Lifestyle Ventures, TMRW, SKORE and TTK Healthcare. On Nautinati, Aditya Birla Fashion's kidswear brand, our images got a 3% click-through rate, against 1.5% for studio shoots and 1% for other AI models.
  • Built the backend and deployment infrastructure on AWS and ShaktiCloud, and tuned inference latency for production.
  • Closed in Dec 2025 when the next round didn't come in time and big labs moved in, turning down ₹28.5 lakh more that would have kept it alive without the quality lead.

Research

Conflict-Aware Adapter Composition (C-AAC)

Computer Vision · Completed

Outperforms base Flux, Imagen 4, and ChatGPT in skin quality and aesthetic appeal.

Read the full write-up →

Flux-VTON+: Hybrid Flux Inpainting and Multi-LoRA Expert Fusion

Generative AI · Completed

State-of-the-art VTON pipeline synergizing Flux Fill and Redux with Multi-LoRA Fusion, achieving an 85% success rate on complex garments.

Read the full write-up →