Overview

Data - AI / ML Engineer

Responsibilities

• We are looking for a Data - AI / ML Engineer who can own both the data pipelines that bring verified information into our platform and the intelligence models that turn that information into reliable startup scores.
• This is a dual-responsibility role. You will be expected to build and maintain robust data infrastructure as well as develop, calibrate, and improve machine learning models. Both are equally important to the platform.
• You will work closely with the Platform Owner and alongside a Backend Engineer who owns system integrations and workflow logic. Your work produces the scored intelligence output. The Backend Engineer's work connects that output to platform actions. The two roles are tightly interdependent and require close daily collaboration, especially in the first 30 days.
• Note: You are the first technical hire on the platform team. The Backend Engineer joins the same week. Clear communication, well-defined handoff points, and shared documentation between the two of you are non-negotiable from Day 1.
• Build and maintain pipelines that collect, clean, and normalize data from multiple external sources into a consistent, usable format
• Design connector architecture that allows individual data sources to be added, swapped, or removed without rebuilding the entire pipeline
• Implement automated data quality checks that catch bad data before it reaches the scoring layer -anomaly detection, constraint enforcement, and schema validation
• Build an automated eligibility screening system that verifies whether a startup has sufficient verified data before assessment begins
• Ensure the pipeline is resilient -critical data signals must have backup sources so a single vendor failure does not disrupt platform output
• Structure data storage to support different regulatory requirements across multiple countries -data from different regions must be handled according to the rules of that region
• Audit the existing scoring engine before making any changes -understand what it does, how it was built, and what would be lost if it were modified
• Calibrate scoring models against real portfolio data so that scores are meaningful, consistent, and comparable across different startup types and stages
• Build confidence scoring logic that determines when the system is certain enough to act autonomously and when it should route to human review
• Ensure every model output is explainable -investors must be able to see exactly which data points drove a score, not just the final number
• Build a feedback loop so that real-world outcomes feed back into the model over time, making it progressively more accurate
• Maintain a structured data store of assessment outputs and outcomes that the model uses to improve
• Define a clear data contract at the handoff point -what data you produce, in what format, and what the Backend Engineer can expect to receive
• Collaborate on trigger logic -what score thresholds or confidence drops should fire what system actions
• Align on data schema requirements so that the APIs the Backend Engineer builds conform to the structure your pipeline produces
• Communicate blockers early -the pipeline and backend system are built simultaneously, so delays on one side directly affect the other
• Document everything you build so the Backend Engineer and Platform Owner can understand, debug, and extend it without depending on you for every question

Qualifications

• Data normalization and schema design
• Automated data quality validation
• API integration across different source types
• Pipeline orchestration and scheduling
• Cloud infrastructure -storage, compute, and deployment
• Version control and code documentation
• AI and Machine Learning
• Building and calibrating supervised machine learning models
• Model explainability -making model outputs traceable and interpretable
• Confidence scoring and threshold calibration
• Experiment tracking and model versioning
• Building feedback loops that improve models over time using real-world outcomes
• Natural language processing or document understanding
• Good to Have
• Vector databases and semantic search
• Good to Have
• Collaboration and Context
• Ability to define clear data contracts and handoff points with backend engineers
• Clear written documentation of pipeline logic, model decisions, and failure modes
• Prior experience working in or with early-stage startups
• Good to Have
• Exposure to financial data, investment platforms, or data verification systems
• Good to Have

Benefits

• Competitive compensation based on experience -discussed during the interview process
• Ownership of both the data and intelligence layers from Day 1 -this is not a support or maintenance role
• Direct access to the Platform Owner and Founder
• Close collaboration with a Backend Engineer from Day 1 -the two roles are designed to work as a unit
• Work on a genuinely novel problem in an emerging market context
• On-site Kolkata with a small, high-accountability team
• Opportunity to scale the platform across multiple international markets