AI Agents & Automation
Agent pipelines that connect systems, tools and LLMs into workflows that run without anyone watching them.
I’m Hetul Shah — an AI and software engineer from Ahmedabad. Computer Science at Nirma University, applied AI as an intern, ten months as a full-time Analyst at Deutsche Bank, and now building AI products of my own through Only Corporate Solutions. A couple of hackathon podiums along the way.
Skills
Three areas I’ve shipped real work in — with the engineering discipline Deutsche Bank drilled into me underneath all of them.
Agent pipelines that connect systems, tools and LLMs into workflows that run without anyone watching them.
Assistants and copilots grounded in your own knowledge base — answers with citations you can actually check.
Turning images, video and scanned documents into structured data a system can act on — at volume.
Achievements
Competitive builds under a clock, judged against everyone else in the room.
Finished runner-up at MINeD, Nirma University’s flagship data and AI hackathon — a multi-day build against a live problem statement, judged on working software rather than a pitch. Scoped the problem down to what could actually be shipped in the time available, then shipped it.
Won first place at zone level in Deutsche Bank’s internal hackathon, competing against teams from across the bank’s engineering org. Built and defended a working solution inside enterprise constraints — real systems, real data-handling rules, no greenfield shortcuts.
Started my own AI practice in 2026 — going from employee to running delivery end to end: finding the problem, architecting the system, building it, and handing it over documented. LeadFlow Pro is the first in-house product out of it.
Selected work
Three that best show the range. More — knowledge assistants, video analytics, document extraction, outreach automation — on request.
My own product — an agent pipeline that discovers, enriches and qualifies B2B leads end to end, no manual step.
Supply chain planning for a pharmaceutical company — solver-driven stock levels and reorder points across warehouses, balancing stockouts against shelf-life expiry.
Real-time visual inspection for a manufacturing line — catches surface defects on parts as they move down the belt.
Experience
Ahmedabad, Gujarat
Run my own AI practice, building custom agents, automation systems, computer vision pipelines and RAG solutions for businesses. I own the whole cycle — scoping the problem on site, architecting, building, and handing over documented systems the client’s own team can maintain. LeadFlow Pro, an autonomous B2B lead generation pipeline, is the first in-house product out of it.
Enterprise systems & financial technology
Built and maintained production systems for financial operations at one of the world’s largest investment banks. This is where I learned what “production-ready” actually costs: code review standards, compliance constraints, release discipline, and system scale where every change touches real money and real customer data. Also took first place at zone level in the bank’s internal hackathon, competing against engineering teams from across the organisation.
Image, video & document intelligence
Worked across the AI division on production-facing projects — image processing pipelines, video analytics, document intelligence and text extraction. Handled the full loop rather than just modelling: data collection, model selection, training, deployment and integration into workflows that people actually used. The internship that turned coursework into shippable systems.
Degree, plus a lot of self-directed AI work
Started building with AI in my second year, well before it was the obvious path — classifiers, computer vision projects, NLP pipelines and deep learning experiments alongside the formal curriculum. Finished runner-up at MINeD, the university’s flagship data and AI hackathon, and graduated with a stack of working projects rather than a stack of certificates.
About
AI & Software Engineer
Computer Science engineer from Nirma University with 4+ years of hands-on AI work — started in college, sharpened it during my AI internship, and got the engineering discipline beaten into me over ten months as a full-time Analyst at Deutsche Bank.
In 2026 I started Only Corporate Solutions to put all three together — academic depth, applied AI, and enterprise rigour — and solve real operational problems for businesses. I still write most of the code myself, and I still prefer a working prototype to a deck.
01 — Depth
Most failed AI projects were framed wrong, not built wrong. I dig until I understand what’s really breaking — the model is usually the easy part once you know that.
02 — Honesty
Some problems need a SQL query and a cron job, not a model. I’d rather say that out loud than ship something impressive that quietly doesn’t work.
03 — Ownership
Documentation, clean architecture and no black boxes. Whoever inherits my code should be able to read it, change it, and not need me on a call to do so.
04 — Measurable
Hours saved, errors removed, throughput gained. I like defining what success looks like before starting — and then actually measuring it afterwards.
Get in touch
Whether it’s a role, a project, or just a problem you’re stuck on — send me a note. I read everything and reply to anything that isn’t a template.
Open to roles, freelance & collaborations · Based in Ahmedabad · Available remotely