Website:
mindpex.com
Job details:
JOB DESCRIPTION
Location: Remote | Duration: 6 Months | Stipend: Zero | Hiring: 2–3 Backend & Applied ML Engineering Interns
THE MISSION — READ THIS OR DON’T BOTHER
We’re building a behavioral intelligence engine that detects invisible attrition risk signals inside companies before employees resign.
Not chatbot AI.
Not another dashboard.
Not “AI” for LinkedIn hype.
We’re building a real backend intelligence system using:
- metadata,
- behavioral signals,
- rule-based inference systems,
- data pipelines,
- and applied ML.
If you enjoy solving messy real-world engineering problems, keep reading.
If you’re looking for a comfortable internship with structured training sessions and predictable tasks, this is probably not for you.
This is startup warfare.
WHAT IS PRHTE?
PRHTE is an attrition detection engine designed to identify hidden employee disengagement patterns using:
- attendance behavior,
- meeting activity,
- communication patterns,
- workflow metadata,
- behavioral volatility,
- and operational signals.
The system combines:
- backend engineering,
- metadata analysis,
- signal engineering,
- rule-based intelligence,
- and practical ML systems.
WE’RE NOT OFFERING:
- Salary
- Perfect datasets
- Clean requirements
- Easy timelines
- Corporate comfort
- “AI research lab” work
WE’RE OFFERING:
- Real product engineering experience
- Exposure to production-grade backend systems
- Applied ML experience with real business problems
- Experience building intelligence systems from scratch
- Portfolio proof that you can actually build
- Ownership from Day 1
- Potential long-term opportunity/equity if things work out
WHAT YOU’LL ACTUALLY DO
No fake intern tasks.
You’ll be:
- Building Python-based data pipelines
- Processing metadata and behavioral logs
- Working with timestamps, attendance patterns, and operational signals
- Extracting and transforming data from APIs
- Building rule-based risk scoring systems
- Engineering behavioral indicators for attrition detection
- Connecting multiple datasets together
- Cleaning messy real-world HR data
- Assisting in practical ML workflows
- Debugging broken pipelines and workflows
- Shipping systems fast under pressure
CORE STACK
MUST KNOW
- Python
- Pandas
- JSON/CSV handling
- APIs & data extraction
- Basic OOP
- Git/GitHub
HUGE ADVANTAGE IF YOU KNOW
- Streamlit
- FastAPI/Flask
- requests/aiohttp
- Cron jobs
- Dashboarding
- Google Workspace APIs
- HRMS systems
- Signal analysis
- Basic ML workflows
IMPORTANT: THIS IS NOT PURE ML RESEARCH
This role is heavily focused on:
- backend systems,
- metadata intelligence,
- behavioral signal engineering,
- practical automation,
- and production execution.
We care more about:
- systems that work,
- than
- fancy models in notebooks.
WHO WE WANT
The Ideal Candidate
Someone who:
- enjoys solving messy problems,
- likes building systems from scratch,
- learns fast independently,
- is comfortable with ambiguity,
- can work under pressure,
- and wants real startup exposure.
WHO SHOULD NOT APPLY
- People looking only for certificates
- Candidates expecting step-by-step spoon-feeding
- Pure theoretical ML learners
- People obsessed with “cutting-edge AI” without understanding systems
- Anyone uncomfortable with messy data and changing requirements
MINIMUM REQUIREMENTS
- Strong Python fundamentals
- Comfortable with Pandas
- Experience handling datasets
- Basic understanding of APIs
- Ability to work independently
- 25–35 hours/week availability
- Strong ownership mindset
SELECTION PROCESS
Step 1 — Screening Questions
We’ll assess:
- Python
- Pandas
- Metadata understanding
- APIs
- Signal thinking
- Problem-solving mindset
- Ownership & communication
Step 2 — 48-Hour Engineering Challenge
You’ll receive:
- attendance.csv
- directory.csv
Task:
Build a Python script that:
- processes metadata,
- calculates attendance volatility,
- identifies behavioral anomalies,
- groups risk by department,
- and explains the top 5 risky employees.
Step 3 — Reality Check
We’ll explain exactly how difficult this will be.
If you’re still excited afterward, you’re probably a fit.
THE HARD TRUTH
This internship will be intense.
You will:
- work with messy data,
- face changing requirements,
- debug unpredictable failures,
- and build under pressure.
BUT YOU’LL ALSO:
- Learn practical engineering extremely fast
- Build real backend intelligence systems
- Understand how production systems actually work
- Gain real startup execution experience
- Build a portfolio that stands out massively
HOW TO APPLY
Send:
- Resume
- GitHub/Portfolio
- A short paragraph on why you want to work on PRHTE
to: [email]
Application Deadline:
When we find the right builders.
"Messy metadata tells the truth before people do."
Come build something difficult with us.
Click on Apply to know more.