Role at a glance
- Position: Senior Director, Data Science (1 position, Panchkula)
- Grade: Senior Director (SD3)
- Reports to: AVP Technology or VP Engineering
- Location: Panchkula: Third Floor, HSIIDC Technology Park, Plot No. 12-A, Sector 22, Panchkula, Haryana 134116 (Chandigarh tri-city). Work from office, with travel as required
- Compensation: INR 30 to 35 Lakhs CTC
- Employment: Full-Time, Permanent
- Experience: 5 to 10 years in data science or machine learning, with models running in production
- Growth path: Senior Director (SD3) > AVP Technology (SD4) > VP Engineering
The mandate
Prozo builds and runs its own technology: ProWMS for warehouse management, ProOMS for orders, ProShip for shipping and courier allocation, and Control Tower for network-wide visibility. The platform handles about 56,000 orders a day across 44 warehouse sites, with peaks above 23,000 API requests a minute. The team releases daily and builds AI-first with Claude.
Prozo already runs machine learning in production. PACE allocates every shipment to a courier partner using gradient-boosted models retrained weekly. NETRA scores every order for RTO risk using models trained on about 13 million orders, and parses every delivery address with a fine-tuned Indian-language model. The Senior Director, Data Science owns these models, builds the next ones, and works with the Founder & CEO on forward-looking projects from the Panchkula office.
Key responsibility areas
1. PACE Courier Allocation
- Own the PACE models (LightGBM) end to end: features, weekly retraining, evaluation and production performance.
- Measure allocation quality against delivery speed,
cost and RTO, and improve the model with the D2C Freight team.
2. NETRA RTO Risk and Address Intelligence
- Improve the NETRA RTO risk models (XGBoost) and the address intelligence model (fine-tuned IndicBERT).
- Explain model outputs in plain terms to brands and operations teams.
3. New Models
- Build volume forecasting, anomaly detection and demand-aware slotting models for the warehouse network.
- Pick problems with the business unit heads where a model changes a decision, and ship them to production.
4. Data and MLOps
- Build the data pipelines, monitoring and retraining jobs that keep every model accurate in production on AWS.
- Work with engineering to capture the data points the models need from ProWMS, ProShip and Control Tower.
5. AI-First Work
- Use Claude and the Anthropic API to build assistants and analysis tools, alongside classical models.
- Use AI coding tools to write, review and test code faster.
Key performance indicators
- Delivery speed, cost and RTO outcomes of PACE allocations against the baseline.
- Accuracy and stability of NETRA models in production.
- New models shipped to production and adopted by the business.
Qualifications & experience
- B.Tech, M.Tech or M.S. in Computer Science,
Statistics, Mathematics or a related field from a Tier 1 or Tier 2 institute.
- 5 to 10 years in data science or machine learning, with models you built running in production.
- Solid Python and SQL, and hands-on experience with gradient boosting (LightGBM, XGBoost) and NLP models.
- Experience with model deployment, monitoring and retraining on a cloud platform, preferably AWS.
- Experience with LLMs and AI coding tools such as Claude Code, Cursor or Copilot.
- Stable tenures. Experience in logistics, e-commerce or marketplace data is a strong plus.
Working at Prozo
- Completely work from office at the Panchkula corporate office (Chandigarh tri-city), with travel as required. Open to candidates based in Panchkula or Chandigarh.
- Monday to Saturday, with the second and fourth Saturdays off.
- You will work with the Founder & CEO, the D2C Freight and warehousing business units, and the engineering team.
- Apply by email to
[email protected] with your resume, current CTC and notice period, and a short note on why this role fits you.
About Prozo
Prozo is India's integrated supply chain operating system, a technology-led 3PL/4PL platform operating 50+ fulfilment centres across 3M+ sq ft, serving 120+ enterprise brands including Tira (Reliance), Pepe Jeans, Bata, Philips, Daikin, Snitch, bewakoof, and others. Our proprietary technology stack, ProWMS, ProShip, ProOMS, and Control Tower, powers omnichannel fulfilment across D2C, marketplace, quick commerce, and B2B channels. Website: www.prozo.com
📌 Senior Director, Data Science (Panchkula)
🏢 Prozo
📍 Panchkula