Data Scientist

Client: S&P Global Market Intelligence (Cognizant)
The Role: Quantitative Research Analyst (Data Modeling & Imputation)
Location: Chicago, IL or Boston, MA (5 days Onsite)

Rate: $70 on C2C or $65 on Hourly W2

The Opportunity

This role focuses on building high-quality, research-ready datasets from incomplete, inconsistent, and fragmented financial data. You will be responsible for designing and implementing imputation frameworks, data transformations, and modeling pipelines that convert raw inputs into reliable signals.

This is not a pure modeling role. The majority of the value comes from data construction, validation, and methodological rigor upstream of the model.

The Impact

You will directly influence the quality and credibility of investment research by:

·         Constructing datasets where ground truth is partial or noisy

·         Designing imputation methodologies that are economically and statistically sound

·         Ensuring outputs are stable, explainable, and usable in production research

Your work underpins signal generation, backtesting, and ultimately client-facing insights.

Core Responsibilities

·         Build and maintain end-to-end data pipelines across structured and unstructured datasets

·         Develop imputation frameworks for missing or sparsely reported financial data (e.g., segment-level estimates, coverage gaps, timing mismatches)

·         Design and implement data normalization and reconciliation logic across overlapping hierarchies (e.g., segments, geographies, entities)

·         Perform data quality diagnostics, including coverage analysis, bias detection, and stability testing

·         Partner with researchers to translate raw data into model-ready features

·         Write efficient, reproducible code in Python and SQL for large-scale data processing

·         Document methodologies clearly to ensure transparency and repeatability

Required Experience

·         5–7 years of experience in quantitative research, data science, or financial data engineering

·         Strong expertise in data wrangling and transformation at scale (this is the core skill)

·         Proven experience with missing data techniques and imputation methods (e.g., cross-sectional inference, time-series interpolation, model-based approaches)

·         Advanced Python skills (pandas, numpy); strong SQL required

·         Experience working with messy, real-world datasets (not just clean academic data)

·         Solid grounding in statistics and econometrics

·         Familiarity with equity markets and financial statements preferred

What We’re Actually Looking For

·         You are an expert data wrangler — you know that 80% of the work is getting the data right

·         You are skeptical of inputs and instinctively test for edge cases, leakage, and bias

·         You understand that imputation is modeling, not a preprocessing afterthought

·         You can balance statistical rigor with practical constraints (coverage vs. precision trade-offs)

·         You write code that others can read, audit, and reuse

Nice to Have (Not Required)

·         Experience with NLP or unstructured data pipelines

·         Exposure to alternative datasets (supply chain, transcripts, etc.)

·         Familiarity with distributed compute environments (Databricks, Spark)