Data engineering · Lead scoring
Lead Generation Data Pipeline in Python and SQL
A Python + SQL pipeline that pulls leads from every campaign channel, cleans and de-duplicates them into one golden record per person, scores them, and reports which campaigns bring leads worth calling.
- In production at Eject Solutions
- 3,000+ qualified leads
By Adithya Reddy Nalla, Lead AI Engineer in Hyderabad.
Problem
Leads arrive from many channels in many formats: Meta lead ads, Google Ads, LinkedIn, and website, CRM and event forms. The same person often appears several times, fields are inconsistent, and without one view nobody can tell which campaigns bring leads worth calling.
Sales teams trust a short list that is reliably good far more than a long one they must filter themselves.
Approach
- Extract: one connector layer (connectors.py) lands every channel in a single raw table, recording each lead's source.
- Clean: Python standardises names, phone numbers and company fields so everything downstream can be trusted.
- De-duplicate: Python and SQL build one golden record per lead.
- Score: SQL scores each lead, and the scores roll up into a campaign_performance report that compares channels on lead quality, not just volume.
Tech stack
- Pipeline
- PythonSQLconnectors.py
- Reporting
- campaign_performance reportPower BI
Results
3,000+
Segmented, qualified leads
₹1.4Cr+
Revenue facilitated within 6 months
4
Channel groups in one pipeline
The leads supported the number one outreach campaign of the year at Eject Solutions.
Architecture
Links
More projects
- Research framework · Agentic AIGBNI: Generative Brand Narrative IntelligenceGBNI (Generative Brand Narrative Intelligence) is a state-aware multimodal AI agent framework for emotionally adaptive brand communication, on SSRN.
- Self-directed R&D · SecurityMITHRA: Cybercrime Rapid-Response and Intelligence PlatformMITHRA is Adithya Reddy Nalla's self-directed R&D platform for responding to cybercrime quickly and turning incident signals into usable intelligence.
- Peer-reviewed · Explainable MLDiabetes Risk Prediction Web AppAn explainable diabetes risk prediction web app: a tuned Random Forest (82%+ accuracy, 0.87 ROC-AUC) with SHAP explanations, deployed on Streamlit.
- Robotics · Embedded AIAglier: Tree-Climbing Fruit-Harvesting RobotAglier is Adithya Reddy Nalla's force-adaptive, lizard-inspired tree-climbing robot that harvests fruit. Sole inventor; the patent specification is in drafting.