DEV Community

FARHAN HABIB FARAZ profile picture

FARHAN HABIB FARAZ

Prompt Engineer focused on conversational AI, voice agents, and automation. Sharing hands-on insights from building and shipping production AI systems.

Location House No. 29, Flat (M-A), Road No. 6, Dhanmondi, Dhaka - 1209, Bangladesh. Joined Joined on  Email address fh.faraz@gmail.com

Education

Premier University, Chattogram

Pronouns

He/Him

Work

Senior Prompt Engineer at PowerinAI

The Silence Problem Nobody Designs For Until A Real Call Goes Quiet

The Silence Problem Nobody Designs For Until A Real Call Goes Quiet

Comments 1
5 min read

Want to connect with FARHAN HABIB FARAZ?

Create an account to connect with FARHAN HABIB FARAZ. You can also sign in below to proceed if you already have an account.

Already have an account? Sign in
Building A Prompt Template That Works Without You In The Room

Building A Prompt Template That Works Without You In The Room

Comments
4 min read
When The Fix Isn't The Prompt At All, It's What The Prompt Is Reading From

When The Fix Isn't The Prompt At All, It's What The Prompt Is Reading From

Comments
4 min read
When A Product Review Shoot Needs You To Be Honest And Complimentary At The Same Time

When A Product Review Shoot Needs You To Be Honest And Complimentary At The Same Time

Comments
4 min read
Reusing A Prompt System Across Clients Without Turning It Into A One Size Fits All Failure

Reusing A Prompt System Across Clients Without Turning It Into A One Size Fits All Failure

Comments
4 min read
Measuring Whether A Training Actually Worked Is Harder Than Delivering It

Measuring Whether A Training Actually Worked Is Harder Than Delivering It

Comments
4 min read
Maintaining A Government Custom GPT After The Training Ends Is A Different Job Than Building It

Maintaining A Government Custom GPT After The Training Ends Is A Different Job Than Building It

1
Comments
4 min read
How Do You Know A Trainer Is Actually Ready, Not Just Trained

How Do You Know A Trainer Is Actually Ready, Not Just Trained

Comments
4 min read
Why Data Privacy And Data Accuracy Are Constantly Fighting Each Other In Client Prompt Systems

Why Data Privacy And Data Accuracy Are Constantly Fighting Each Other In Client Prompt Systems

Comments
4 min read
Leading A Prompt Team Means Debugging People's Mental Models, Not Just Their Prompts

Leading A Prompt Team Means Debugging People's Mental Models, Not Just Their Prompts

Comments
4 min read
Building A Training Module Is A Prompting Problem Long Before It Is A Content Problem

Building A Training Module Is A Prompting Problem Long Before It Is A Content Problem

Comments
4 min read
Directing Video Content Nobody Is Actually Behind A Camera For

Directing Video Content Nobody Is Actually Behind A Camera For

Comments
4 min read
What Journalists Taught Me About Writing Prompts For Skeptics

What Journalists Taught Me About Writing Prompts For Skeptics

Comments
4 min read
Concept Notes And Pitch Decks Are Not The Same Document Wearing Two Outfits (Even Though Everyone Treats Them That Way)

Concept Notes And Pitch Decks Are Not The Same Document Wearing Two Outfits (Even Though Everyone Treats Them That Way)

Comments
4 min read
How A Product Review Shoot Taught Me More About Prompting Than Most Prompting Work Did

How A Product Review Shoot Taught Me More About Prompting Than Most Prompting Work Did

Comments
4 min read
What Happens When You Have To Teach Someone Else To Do Your Job (And They Cannot See What You See)

What Happens When You Have To Teach Someone Else To Do Your Job (And They Cannot See What You See)

Comments
3 min read
The Custom GPT That Had To Behave Like A Government Employee, Not A Chatbot

The Custom GPT That Had To Behave Like A Government Employee, Not A Chatbot

Comments
3 min read
Why Tender Documents Break Most AI Writing Tools (And What Actually Works)

Why Tender Documents Break Most AI Writing Tools (And What Actually Works)

Comments
3 min read
The Middleman Problem: Why Fixing AI Chatbots Is Never Just About the Prompt

The Middleman Problem: Why Fixing AI Chatbots Is Never Just About the Prompt

Comments
3 min read
The RAG System That Found Contradicting Answers (And Confidently Picked The Wrong One)

The RAG System That Found Contradicting Answers (And Confidently Picked The Wrong One)

Comments
4 min read
The RAG System That Mixed Documentation From Different Products (And Created Frankenstein Instructions)

The RAG System That Mixed Documentation From Different Products (And Created Frankenstein Instructions)

Comments
5 min read
The RAG System That Retrieved Perfect Chunks (But Answered Wrong Anyway)

The RAG System That Retrieved Perfect Chunks (But Answered Wrong Anyway)

Comments
4 min read
The Calendar Sync That Scheduled Meetings During Weekends (And National Holidays)

The Calendar Sync That Scheduled Meetings During Weekends (And National Holidays)

Comments 1
4 min read
The Feedback Collector That Published Negative Reviews Publicly (Before Human Review)

The Feedback Collector That Published Negative Reviews Publicly (Before Human Review)

Comments
4 min read
The Product Recommender That Only Suggested Out-of-Stock Items (Inventory Integration Fail)

The Product Recommender That Only Suggested Out-of-Stock Items (Inventory Integration Fail)

Comments
4 min read
The Onboarding Flow That Ghosted 300 New Users (Missing Conditional Logic)

The Onboarding Flow That Ghosted 300 New Users (Missing Conditional Logic)

Comments
4 min read
The Notification System That Sent 12,000 Messages at 3 AM

The Notification System That Sent 12,000 Messages at 3 AM

Comments
4 min read
The Translation Bot That Turned Professional Emails Into Insults

The Translation Bot That Turned Professional Emails Into Insults

Comments
4 min read
The FAQ Bot That Made Up Answers When It Couldn’t Find Real Ones

The FAQ Bot That Made Up Answers When It Couldn’t Find Real Ones

Comments
4 min read
The AI That Quoted Customers Their Competitors’ Prices

The AI That Quoted Customers Their Competitors’ Prices

Comments 1
3 min read
The Booking System That Created 47 Double-Bookings in One Morning

The Booking System That Created 47 Double-Bookings in One Morning

Comments
4 min read
The Email Responder That Replied Unsubscribe to the CEO

The Email Responder That Replied Unsubscribe to the CEO

Comments
3 min read
The Sentiment Analysis Bot That Thought Every Customer Was Angry

The Sentiment Analysis Bot That Thought Every Customer Was Angry

Comments
3 min read
The Lead Magnet That Collected 5,000 Fake Emails And We Didn’t Notice for 2 Weeks

The Lead Magnet That Collected 5,000 Fake Emails And We Didn’t Notice for 2 Weeks

Comments
5 min read
The Chatbot That Started Arguing With Customers Because I Told It to Be Confident

The Chatbot That Started Arguing With Customers Because I Told It to Be Confident

Comments 1
6 min read
The Refund Bot That Approved $127,000 in One Afternoon

The Refund Bot That Approved $127,000 in One Afternoon

Comments
4 min read
The Appointment Scheduler That Booked Everyone for Yesterday

The Appointment Scheduler That Booked Everyone for Yesterday

1
Comments
4 min read
The Chatbot That Forgot English Mid-Conversation

The Chatbot That Forgot English Mid-Conversation

Comments
4 min read
The Medical Chatbot That Diagnosed Everyone With Cancer

The Medical Chatbot That Diagnosed Everyone With Cancer

Comments 2
4 min read
The Invoice That Cost $470,000 (Because I Forgot One Word)

The Invoice That Cost $470,000 (Because I Forgot One Word)

Comments
3 min read
The Workflow That Worked Too Well (And Crashed Everything)

The Workflow That Worked Too Well (And Crashed Everything)

Comments
4 min read
The Day I Accidentally Made Our AI Sound Like a Desperate Salesperson

The Day I Accidentally Made Our AI Sound Like a Desperate Salesperson

Comments
4 min read
The Time Our Voice AI Couldn’t Understand “Yes” (And the Cultural Lesson That Fixed It)

The Time Our Voice AI Couldn’t Understand “Yes” (And the Cultural Lesson That Fixed It)

Comments
4 min read
The Knowledge Base That Lied to 10,000 Customers (And How We Caught It)

The Knowledge Base That Lied to 10,000 Customers (And How We Caught It)

Comments
6 min read
The Day Our CRM Automation Started Flirting With Customers

The Day Our CRM Automation Started Flirting With Customers

Comments
4 min read
The “Too Smart” Knowledge Base Problem: When Your AI Knows Too Much for Its Own Good

The “Too Smart” Knowledge Base Problem: When Your AI Knows Too Much for Its Own Good

Comments
5 min read
Sentiment-Aware Response Tuning: When the Same Question Needs Different Answers

Sentiment-Aware Response Tuning: When the Same Question Needs Different Answers

Comments
6 min read
Conversation Flow Control: When Users Don’t Follow Your Script

Conversation Flow Control: When Users Don’t Follow Your Script

Comments
5 min read
The Voice AI Prompt That Was Too Human (And Why Robot Instructions Worked Better)

The Voice AI Prompt That Was Too Human (And Why Robot Instructions Worked Better)

Comments
4 min read
Multi-Channel Chatbot Synchronization: When Your Bot Has Multiple Personalities Across Platforms

Multi-Channel Chatbot Synchronization: When Your Bot Has Multiple Personalities Across Platforms

Comments 2
6 min read
The RAG System That Confidently Made Up Product Specs (For 6 Months Straight)

The RAG System That Confidently Made Up Product Specs (For 6 Months Straight)

Comments
7 min read
The Insurance Bot With 47 Decision Points (And How It Approved Everything By Accident)

The Insurance Bot With 47 Decision Points (And How It Approved Everything By Accident)

Comments 1
7 min read
The Lead Scoring System That Ignored Million-Dollar Deals (Because The CEO Had a Gmail Account)

The Lead Scoring System That Ignored Million-Dollar Deals (Because The CEO Had a Gmail Account)

Comments
7 min read
The Pricing Bot That Quoted $847 for a $12 Item (When Complexity Goes Wrong)

The Pricing Bot That Quoted $847 for a $12 Item (When Complexity Goes Wrong)

Comments
7 min read
The Day Users Discovered They Could Hack Our Chatbot With "Ignore Previous Instructions"

The Day Users Discovered They Could Hack Our Chatbot With "Ignore Previous Instructions"

Comments 1
6 min read
The Voice AI That Thought "Thursday" Was "Birthday" (Accent Recognition Gone Wrong)

The Voice AI That Thought "Thursday" Was "Birthday" (Accent Recognition Gone Wrong)

Comments
6 min read
The Client Who Wanted AI to "Remember Everything" (And Why That Was a Terrible Idea)

The Client Who Wanted AI to "Remember Everything" (And Why That Was a Terrible Idea)

Comments
5 min read
The Chatbot That Couldn't Understand "Hmm" (Until We Taught It Bangla Body Language)

The Chatbot That Couldn't Understand "Hmm" (Until We Taught It Bangla Body Language)

Comments
4 min read
01670-309-328 vs +8801670309328: How We Solved a Western Consulting Firm's Phone Number Nightmare

01670-309-328 vs +8801670309328: How We Solved a Western Consulting Firm's Phone Number Nightmare

3
Comments
3 min read
Conversation Memory Collapse: Why Excessive Context Weakens AI

Conversation Memory Collapse: Why Excessive Context Weakens AI

Comments
3 min read
loading...