Irfan Ali — Senior AI Engineer

Senior AI engineer building LLMs, RAG, and agents in production.

I build production LLM systems end to end — RAG that returns the right answers, voice agents on real telephony, and multi-agent pipelines that hold up under load. Most recently: a B2B SaaS live with a paying customer, built solo.

Open to full-time senior AI engineer roles — remote, based in India, working with global teams.

Seven-plus years building AI and data systems. Previously founding AI engineer at Kuration AI in Hong Kong, and the first AI hire at Schneider Electric in India, reporting to the CTO.

Download résumé

Get in touch

View open source

datacortex.summary

$datacortex --info

WHOIrfan Ali

WHATLLMs · RAG · Voice

YEARS7+

LIBS12 PyPI

LOCRemote / India

Open to full-time roles

ProductsGodamPaying customer StacksiftB2B SaaS ReflectaVoice AI AgentEnsembleAgents ragfallbackRAGConnect

LinkedIn GitHub PyPI

Explore

products.shipped

$ls ./production

Shipped Products

Production systems I designed and built end to end — from billing SaaS with a paying customer to voice AI and B2B analytics.

getgodam.com

Godam

Live · Paying customer

B2B inventory & billing SaaS for FMCG distribution in Nepal. Runs the client's daily billing, payments, credit ledgers, and field-sales logging.

Next.js 15TypeScriptSupabase

Visit getgodam.com

Stacksift

Live payments

B2B domain product analyzer running a 5-stage LLM classification pipeline (DSPy + GPT-4.1) at ~$0.03 per analysis — ~84% gross margin.

FastAPIDSPyRazorpay

stacksift.in →

Reflecta

Solo build · In production

Voice-first AI wellness app. Phone check-ins over telephony with post-call analysis extracting 35+ wellness signals per call — burnout risk scoring, crisis-keyword detection — plus pgvector memory and multi-provider LLM fallback.

FastAPIBolnaGroqpgvector

getreflecta.com →

pypi.org/user/irfanalidv

$pypi list --user irfanalidv

> 12 libraries on PyPI — free for other engineers building AI systems.

RAGNav

Hybrid BM25 + dense retrieval with RRF fusion. Navigation-first RAG for long documents. R@3 = 0.956 on SQuAD · 131 tests.

ragfallback

Stop RAG from failing silently. Query rewriting, retrieval confidence scoring, fallback strategies, retry logic.

AgentEnsemble

Multi-agent orchestration. ReAct, Swarm, Pipeline, Debate, WorkflowGraph patterns. Routing, planning, RAG, cost tracking.

nepal-gov-agent

Agentic RAG on Nepal government policy and legal documents. Hybrid retrieval with citations, offline-capable, Nepali + English.

AgentCare

Voice AI for healthcare. Call intake, structured extraction, missing-data recovery, appointment orchestration, post-call analytics.

scrapeflow-py

Playwright-based web scraping. LLM extraction, hybrid selectors, session persistence, rate limiting, anti-detection.

AskPandas

Natural-language queries on CSV data using local LLMs. No API keys, no cloud.

lingo-nlp-toolkit

Lightweight NLP utilities bridging traditional pipelines and transformer-ready workflows.

PyroChain

Agentic feature engineering. PyTorch + LangChain agents for multimodal feature extraction.

toxic-comment-classifier

Deep learning toxicity detection. Per-category scores for obscene language, threats, insults, identity hate.

socialmediaextractor

Extract social media profile links from websites. Useful for lead enrichment and web intelligence pipelines.

trustpilot-scraper

Scrape Trustpilot reviews programmatically. Structured output for sentiment analysis and competitive research.

Also: Stacksift — a B2B domain analyzer built on these patterns. stacksift.in →

View all on GitHub View all on PyPI

what-i-build

$what-i-build

What I build

Production LLM systems — the kind teams rely on after launch, not just in demos. Seven-plus years across startups and enterprise R&D.

RAG systems

Retrieval pipelines that return the right answers — query rewriting, confidence scoring, fallback strategies, and navigation-first retrieval for long documents.

Examples: ragfallback, RAGNav

Agent orchestration

Multi-agent pipelines with ReAct, swarm, pipeline, debate, and workflow patterns. Routing, planning, and cost tracking built for production.

Examples: AgentEnsemble

Voice AI

Phone and in-app voice features — call intake, structured extraction, post-call LLM analysis, and appointment orchestration.

Examples: Reflecta, AgentCare

Multi-provider LLM orchestration

Systems that route across providers, fall back when one fails, and stay reliable under real load — not just demo prompts.

Examples: Production patterns across shipped products

Hiring for a senior AI role? Let's talk.

Get in touch

experience

$resume --experience

> 7+ years building AI and data systems for startups and enterprise

ROLES

Independent
Senior AI Engineer
Built and sold Godam, a B2B inventory & billing SaaS live in production with a paying customer in Nepal. Also shipped Stacksift (B2B domain analyzer, live payments) and Reflecta (voice-first AI wellness app). Maintains 12 open-source Python libraries on PyPI. Open to full-time senior AI engineering roles.
Kuration AI
Founding AI Engineer
First AI hire at an early-stage B2B startup. Owned the full AI architecture — lead-generation agent pipelines, a web intelligence layer with LLM extraction and Playwright dynamic browsing, and multi-provider LLM orchestration with cost tracking.
Luminous Power Technologies (Schneider Electric)
Senior Manager, Data & Analytics, R&D
Sole AI hire reporting to the CTO. Built the R&D data and AI function from scratch, shipped an LLM-powered dealer assistant (GPT-4 plus a domain fine-tuned model), and hired and led a team of three.
Lynk
Data Analytics & Automation
Analytics and NLP-powered search for an expert-matching platform.
brainsfeed
Head of Data & Analytics
First data hire. Grew a distributed team of 10+ and shipped Infosphere, an intelligence platform with 15+ NLP-enriched attributes and natural-language search.

BUILT_FOR

AI search and chatbots that need to be reliable
Voice AI on the phone or in apps
Multi-step AI agents that hold up in production
Data systems for startups and enterprise R&D

India · Previously Hong Kong

about

$whoami

> AI engineer · 12 open-source libraries on PyPI · two published papers

founder.png

Irfan AliBuilding AI that works in production

I_FOCUS_ON

Reliabilityoverhype

Systemsoverscripts

Maintainabilityovershort-term hacks

NAME

Irfan Ali

EDUCATION

M.Sc. Data Science & AI, IISER Tirupati · B.Tech CSE, Alliance University · Exchange semester at ISEP, Paris

BACKGROUND

Built Godam — a B2B SaaS live in production with a paying customer — plus Stacksift and Reflecta. Founding AI engineer at Kuration AI in Hong Kong. First AI hire at Schneider Electric in India, reporting to the CTO. Open to full-time senior AI engineering roles.

-12 open-source Python libraries on PyPI that other engineers use to build AI
-Built systems that switch between AI providers automatically when one fails
-Two peer-reviewed papers published in 2025

EXPERIENCE

Led data and AI teams across India and Hong Kong
Built AI systems that real businesses use to make real decisions

datacortex.contact

$datacortex contact

> Have a question or want to get in touch? Drop a message below.

formSend a message

Tell me what you're working on or what you'd like to discuss.

Quick & Secure

Your information is protected and will only be used to respond to your inquiry.

Privacy Protected • I'll respond within 24 hours

$other ways to reach me

Other ways to reach me

Email or social — whichever is easier.

irfan.ali@datacortex.in LinkedIn GitHub

What I work on

LLM systemsRAG pipelinesAgent orchestrationVoice AIOpen source

What to expect

1Reply within 24 hours on weekdays

2Direct, async-first conversation