Skip to content
View 25sheraz's full-sized avatar

Achievements

Achievement: Pull SharkAchievement: YOLO

Achievements

Achievement: Pull SharkAchievement: YOLO

Highlights

Block or report 25sheraz

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
25sheraz/README.md

Sheraz Ahmad

AI Engineer · Agentic AI · RAG · AI Automation
I build AI agents, RAG systems and WhatsApp automations that businesses use every day.

Portfolio LinkedIn Email CV (PDF)

📍 Lahore, Pakistan · Open to AI Engineer and AI Automation Engineer roles, remote or on-site

What I've delivered in production

Result What I built Stack
+30% revenue in the first month GoHighLevel CRM lead generation (DevelMO): B2B leads pulled from Apollo, cleaned and stored in GoHighLevel, AI-personalised emails with automated follow-ups, sentiment analysis on every reply, and a meeting booked for each positive one GoHighLevel, Apollo, LLMs
95% accuracy understanding callers AI voice cold-calling agent (DevelMO): pulls each lead from the CRM and opens the call with a personalised message. Now sold as a DevelMO product ElevenLabs, CRM API
24/7 sales on WhatsApp CarBatterySquad WhatsApp sales & support bot (Dubai): handles battery and jump-start enquiries from the first message to a closed sale, remembers returning customers, alerts the team after every sale or complaint, and lets staff take over any chat n8n, WhatsApp Cloud API, LLM, PostgreSQL
Instant, business-specific answers 24/7 Neoline WhatsApp customer service bot: answers any question about Neoline's services from its own service information and FAQs using RAG, so customers get accurate replies without waiting for staff n8n, WhatsApp Cloud API, Qdrant, OpenAI embeddings
500+ queries a month, 90% solved without staff AI support chatbots for Sylink and SecureTapVPN that answer product and FAQ questions from each company's own information, remember the conversation, and hand off to a live agent when needed n8n, DeepSeek, webhooks
90% accuracy on clear meal photos FoodMate nutrition backend: a vision-language model names the meal and estimates calories and macros, plus an AI nutrition coach VLM, REST API
12 hours a week saved, stronger brand engagement LinkedIn & SEO blog automation: a LinkedIn workflow for Tecclub Technology that publishes posts and sends outreach messages, plus a blog pipeline that researches each topic with live web search, writes an 1,800-word SEO article and publishes it to the CMS n8n, OpenAI, LinkedIn API, web search
25 hours a week saved Business workflow automation (Tecclub Technology): n8n workflows that connect ML models to business systems through REST APIs and webhooks n8n, REST APIs, webhooks

Client code is private. I'm happy to walk through the architecture of any of these in an interview.

Open-source projects

Project What it does Stack
n8n AI Automations Production workflows with setup guides: a two-agent WhatsApp sales bot (97% of 200 test scenarios correct), an SEO blog pipeline, an ElevenLabs outbound calling agent, website support chatbots and a vision-based nutrition API n8n, Claude Sonnet 4.6, ElevenLabs, GPT-5.4
Self-Learning Coding Agent Writes Python, runs it in a sandbox, fixes itself from the traceback and saves each failure as a lesson for next time. HumanEval pass@1: 62% → 74% with memory on LangChain, Kimi K2.7-Code, Python
Hybrid RAG Assistant Dense + BM25 search in Qdrant merged with Reciprocal Rank Fusion, cross-encoder reranking, Redis chat memory, a document management API and a React chat widget FastAPI, Qdrant, Redis, React
Pakistan Penal Code Legal Assistant Describe an incident and get the applicable PPC sections, punishments and your rights, with exact citations. OCR fallback for scanned pages GPT-4o-mini, PyMuPDF, RapidOCR, Flask
Optivortium Alliance: AI Eyewear Platform Final-year project: real-time AR virtual try-on using a 478-point facial landmark model, plus a face shape classifier with 87% accuracy on 700+ test images, served via a Flask REST API to recommend glasses MediaPipe, TensorFlow, scikit-learn, Flask, Firebase
Todo App + MCP Server Built by AI coding agents from Linear tickets, with automated tests, 4 GitHub Actions workflows and a custom MCP server that lets LLMs answer questions about the codebase JavaScript, MCP, GitHub Actions
3D Portfolio · live A WebGL "agent graph" that changes shape as you scroll through my work Three.js, GSAP, Vite

Tech stack

Area Tools
Agents & LLMs LangGraph, LangChain, LangSmith, OpenAI API, tool calling, multi-agent workflows, VLMs
RAG & search Qdrant, hybrid search (dense + BM25), reranking, embeddings, ChromaDB
Automation n8n, WhatsApp Cloud API, LinkedIn API, GoHighLevel, ElevenLabs, Apollo, Zapier, webhooks
Backend & data Python, FastAPI, Flask, Node.js, React, PostgreSQL, Redis, MongoDB, Docker
ML & computer vision TensorFlow, scikit-learn, OpenCV, MediaPipe, CNNs, fine-tuning

Experience

  • Generative AI & Automation Engineer · DevelMO · Mar 2026 – Oct 2026
  • AI / ML Developer · Tecclub Technology · Feb 2025 – Jan 2026
  • AI & Deep Learning Intern · Fenris Group · Apr 2023 – Jun 2023
  • BS Software Engineering · The University of Faisalabad · 2026 · CGPA 3.61 / 4.00

Hiring for an AI or automation role? Email sherazseo51@gmail.com or message me on LinkedIn.

Pinned Loading

  1. hybrid-rag-assistant hybrid-rag-assistant Public

    Production-style RAG system: dense + BM25 hybrid search in Qdrant with RRF fusion, cross-encoder reranking, Redis chat memory, a FastAPI document-management API and a React chat widget.

    Python

  2. ppc-legal-rag-chatbot ppc-legal-rag-chatbot Public

    AI legal assistant for the Pakistan Penal Code (PPC 1860). Describe an incident to get the applicable sections, punishments and your rights, with exact citations. RAG, GPT-4o-mini and OCR.

    Python

  3. Self-Learning-Agent Self-Learning-Agent Public

    Autonomous coding agent that writes Python, runs it in a sandbox, fixes its own errors and learns from every failure. HumanEval pass@1 rose from 62% to 74% with self-learning memory.

    Python

  4. n8n-ai-automations n8n-ai-automations Public

    Production n8n workflows: a two-agent WhatsApp sales bot (97% on 200 test scenarios), an SEO blog pipeline, an ElevenLabs calling agent, VPN support chatbots and a vision nutrition API. Cleaned of …

  5. DevelMO DevelMO Public

    DevelMO AI Knowledge Assistant: hybrid RAG chatbot with dense + BM25 search in Qdrant, RRF fusion, cross-encoder reranking, Redis session memory, a FastAPI server and a React chat widget.

    Python

  6. ai-eyewear-platform ai-eyewear-platform Public

    Final-year project: real-time AR virtual try-on with a 478-point facial landmark model, and a face shape classifier (87% on 700+ test images) served via a Flask REST API to recommend glasses.

    HTML