Irfan Ali — AI Engineer

Experience

1. AI Engineer · DataCortex

   Dec 2025 – Present

   India

   Built the AI products end to end, from architecture to deployment and evaluation: a tool that reads a company's website and works out what products they sell (flagging uncertain results for human review), an evaluation set of companies with known answers to measure how accurate it really is, a system that handles check-in calls and writes up what was said, and a stock and billing app that a distributor in Nepal uses every day. Also published open-source Python libraries.

2. Founding AI Engineer · Kuration AI

   Jun 2024 – Nov 2025

   Hong Kong

   Joined early and owned the AI side of the product, taking the company research platform from nothing to customers using it daily. Built a research tool that filled in missing company data from web search and the companies' own sites, stopping once it had enough evidence; a reliability layer that returned a controlled result instead of a wrong one and switched providers when one failed; monitoring that traced a failure to the exact step; a hand-checked set of companies with known answers that the team released against, so a change that scored worse did not ship; and a fit score that said "not qualified" when evidence was thin instead of guessing.

3. Senior Manager (R&D) · Luminous Power Technologies (a Schneider Electric company)

   Aug 2023 – Jun 2024

   India

   Set up the company's AI engineering function, reporting to the CTO. Built Luminous Genie, an assistant that routed requests across business and engineering workflows, and Eagle Eye, a sizing tool where the numbers came from fixed engineering rules rather than the model guessing, plus an internal dashboard for geospatial data and live monitoring on Azure. Hired and led a team of three delivering AI products across sales, R&D and manufacturing.

4. Data Analytics & Automation Associate · Lynk

   Feb 2022 – Jul 2023

   India

   Improved search so clients found the right expert faster. Built the data pipelines, internal APIs and reporting behind it. Fully remote.

5. Head of Data & Analytics · brainsfeed

   Aug 2018 – Feb 2022

   Hong Kong

   Set up the company's data function from scratch and set its working standards. Built systems for understanding and recommending content. Hired and led a team of ten across several countries, fully remote.

Skills and tools

Languages

Python · TypeScript · SQL

AI / LLM

OpenAI · Anthropic · Google Gemini · Groq · One interface across providers · Hugging Face · PyTorch · RAG (keyword and meaning-based search) · Agents and tool use · MCP · Structured output and guardrails · Voice AI · Hand-checked evaluation sets · Confidence scoring and human review · LangSmith tracing

Backend and data

FastAPI · Pydantic · Webhooks and background jobs · Postgres · pgvector · MongoDB · Redis · Supabase · Docker · CI/CD

Frontend and cloud

Next.js · Vercel · AWS · Google Cloud · Azure

Education

Master's in Data Science & AI, IISER Tirupati (2026, CGPA 8.9/10) · B.Tech CSE, Alliance University · International exchange in Data Science & Big Data Analytics, ISEP Paris

Book: Data Voyage: Building Real AI Systems from Data to Deployment (2026, 377 pages, code on GitHub)

Papers:

• Advanced Cross-Validation Framework for Mental Health AI: BERT and Neural Networks Achieve High Accuracy on MentalChat16K
• Multi-Aspect Temporal Topic Evolution with Neural-Symbolic Fusion and Information Extraction for Yelp Review Analysis: A Comprehensive Deep Learning Framework

ORCID 0000-0003-0022-3047.

Languages: English (full professional) · Hindi (native) · Bengali (native) · Nepali (working)

Selected work

• Stacksift — Paste a company website. Get the products they actually sell — with links to the pages that prove it.
• Reflecta — A three-minute phone check-in that turns a week into a short picture for therapy.
• Godam — Billing, stock, and field visits for FMCG distributors in Nepal — built for a phone.

Open source

Published Python libraries, including AgentEnsemble, ragfallback, AgentCare, RAGNav. Full list.

How I work

• -I write things down. Most of my work happens async.
• -I check whether something works before I show it off.
• -I reply within a day on weekdays.

This is my personal site. DataCortex IQ is just the entity the domain sits under.

Hiring FAQ

Are you looking for a job?

Yes. Full-time AI engineering on a product team.

Where are you based?

Bengaluru, India. Remote is fine.

When can you start?

Immediately.

What do you want to own?

Building AI tools and the checks that show whether they are actually working.