AI/ML Engineer with a track record of taking ideas from zero to production.
I've built products serving hundreds of daily users, AI systems adopted by
40+ newsrooms, and open-source infrastructure for LLM memory. I specialize
in building scalable AI applications that deliver measurable business outcomes
not research demos.
work_history
Case Studies
call
Voice AI Receptionist
stack Vapi · Google Calendar · Airtable
Automate call scheduling with a voice AI receptionist. Answers calls, checks calendar availability, books appointments, sends confirmations, and logs everything to Airtable.
Read full case study arrow_forward
search
Real-Time Job Scraper
stack Camoufox · Playwright · Anti-Bot Evasion
Bypass anti-bot detection to scrape 1,000+ job listings weekly from 7 ATS platforms. Uses Camoufox stealth browser with automatic duplicate detection.
Read full case study arrow_forward
reviews
Google Review Automation
stack n8n (No-Code) · Google Business Profile · OpenAI
Automated Google review management: AI-generated responses, sentiment classification, Slack alerts for negative reviews, and full logging to Google Sheets.
Read full case study arrow_forward
1 / 3
star
Testimonials
Arman Liaghat - CEO & Founder, thehired.ai
folder_open
Projects
Leadork
trending_up 100+ daily active users
A B2B lead scraping and prospecting platform. Input any location
and business type anywhere in the world, and Leadork automatically
finds companies, extracts decision-maker emails, and delivers
ready-to-use sales sheets. Used daily by sales teams and agencies
across multiple markets.
AI-powered document verification engine for investigative journalism.
Upload a document and get a forensic analysis report in minutes —
tamper detection, metadata tracing, archive cross-referencing, and
provenance tracking. Built to give African newsrooms the tools to
verify documents and publish with confidence.
A graph-enabled RAG system that gives LLMs persistent memory.
Extracts entities and relationships from conversations and stores
them as a dynamic knowledge graph, enabling long-term context
without context window limits. Designed for applications that
need to remember what was said.
A tool that processes and analyzes audiovisual communication data
from driving simulator experiments. Answers two core questions:
"Who spoke when?" and "What was said?", using speaker diarization
and transcription. Provides visualizations for speaker timelines,
sentiment distribution, and interaction mapping.