investment-data-infrastructure

Chartered Financial Intelligence Architect (CFIA) · 46 questions

• An HFT firm trains an ML-based execution model on two years of historical equity data sourced from the Securities Information Processor (SIP) consolidated feed, which aggregates best-bid/best-offer quotes across all exchanges with approximately 2–5 ms latency. The firm then deploys this model in production against a direct proprietary exchange feed with sub-100 microsecond latency. After deployment, the quant team observes that the model's input feature distributions at inference time diverge substantially from those observed during training, and realized alpha is 60% lower than backtest projections. What is the PRIMARY cause of this performance degradation?
• A quantitative research team must store and efficiently query five years of Level 2 order book data for 3,000 equity symbols. Each record captures the bid and ask queues at 10 price levels with nanosecond-precision timestamps, generating approximately 4 TB of raw data per year. The dominant query pattern involves three operations: (1) time-range slices by symbol and date, (2) point-in-time order book reconstruction at arbitrary timestamps, and (3) aggregation of volume-weighted metrics across trading sessions. The infrastructure team proposes four candidate architectures. Which is MOST appropriate?
• Following a cross-border merger, a global asset manager discovers that the same corporate bond is identified by three different identifiers across its systems: CUSIP in the North American trading desk system, ISIN in the European risk platform, and a proprietary internal code in the valuation engine. Each system also carries a different day-count convention for interest accrual — Actual/360, Actual/365, and 30/360, respectively. A reconciliation report flags a 12-basis-point pricing discrepancy on a €500M position. Tracing the discrepancy, the data team confirms all three systems are receiving the same vendor price feed but computing accrued interest differently. The root cause is BEST described as:
• A long/short equity portfolio manager wants to incorporate satellite imagery of retail parking lots as an alpha signal to predict same-store sales results before quarterly earnings announcements. The data science team delivers a parking lot occupancy index that achieves a Sharpe ratio of 1.4 and an information coefficient of 0.18 in a five-year backtest. Before approving live deployment, the CIO raises a concern about 'backtest performance driven by selection bias in the satellite coverage universe.' Which risk does the CIO MOST likely have in mind?
• An investment research platform is building a knowledge graph that links entities across six node types: companies, executives, suppliers, customers, patents, and regulatory filings. A semiconductor analyst asks the platform to identify 'all companies with greater than 30% revenue exposure to a Tier-1 foundry if that foundry experiences a production halt lasting 90 days, including second-order effects through contract manufacturers that source from that foundry.' Why does a knowledge graph handle this query more naturally than a traditional normalized relational database schema with the same underlying data?
• During a post-stress-event regulatory examination, a buy-side risk manager is asked by examiners to provide full provenance for a VaR figure that appeared in a stress report filed 18 months earlier. The examiners specifically request documentation of: (1) the exact raw market data inputs used on that date, (2) each data cleaning and transformation step applied, (3) which model version produced the VaR output, and (4) the compute environment in which the calculation ran. The firm's infrastructure logs model version tags and compute job IDs but does not track how raw market data was transformed before entering the risk engine. What critical data infrastructure capability is the firm MISSING?
• An AI governance committee at a mid-size asset management firm approves the deployment of an NLP sentiment model that scores corporate earnings call transcripts on a scale of –1.0 to +1.0 to generate a management credibility signal. Six months into live deployment, a compliance audit reveals that the model was trained on a dataset of transcripts supplied by a third-party vendor in which all forward-looking statements containing negative qualifiers (e.g., 'risk,' 'uncertainty,' 'decline') had been redacted to satisfy the vendor's content licensing restrictions. The model's training corpus systematically underrepresents negative management sentiment. This situation most directly illustrates a failure in which governance control?
• A quantitative investment firm's ML team builds a three-factor return prediction model using features derived from price momentum, earnings revision signals, and NLP-scored news sentiment. The model achieves an annualized information ratio of 0.85 in backtesting. After eight weeks of live trading, the team discovers that the sentiment feature in the production serving pipeline uses a 1-hour rolling lookback window to compute a moving-average sentiment score, while the offline training pipeline used a 24-hour rolling lookback window for the same feature. The realized information ratio drops to 0.22 in live trading. This is the canonical example of which ML infrastructure failure?
• A systematic equity fund ingests price data from six different vendors — Bloomberg, Refinitiv, FactSet, ICE, Quandl, and a proprietary exchange feed — each using different corporate action conventions, dividend adjustment methodologies, and ticker symbology. A quant analyst discovers that backtested returns diverge by as much as 340 basis points across vendor sources for the same strategy. What is the most operationally robust normalization approach to eliminate this cross-vendor drift?
• A high-frequency trading desk needs to store and query 10 years of tick-by-tick trade and quote data for 8,000 US equities, averaging 2.5 billion ticks per trading day. The primary access pattern is time-series range queries by symbol and timestamp for backtesting and signal research. Which storage architecture best satisfies throughput, compression, and query performance requirements for this workload?
• A fundamental long/short equity fund is evaluating a new alternative data product: anonymized consumer credit card transaction data aggregated by merchant category. Before signing a data license and integrating the feed into its research process, the Chief Data Officer must conduct due diligence. Which vetting dimension most directly determines whether the dataset creates material legal and reputational risk for the fund?
• A mid-sized asset manager is designing its next-generation cloud data platform to support three use cases simultaneously: (1) raw alternative data ingestion and exploration by quant researchers, (2) structured SQL analytics for performance reporting, and (3) training ML models on multi-year historical datasets. The CTO must choose between a pure data warehouse, a pure data lake, or a hybrid lakehouse architecture. Which architecture best serves all three use cases?
• During an SEC examination, regulators request documentation demonstrating exactly how the NAV for a fixed-income fund was calculated on a specific date six months ago, including the source of each bond's price, any adjustments applied, and the identities of systems that transformed the data before it reached the portfolio accounting system. Which data management capability most directly satisfies this regulatory request?
• A global investment bank's trading desk discovers that risk reports and front-office systems are using conflicting GICS sector classifications for 340 securities, because three different data vendors assign sector codes differently for the same issuers. As a result, the firm's sector-neutral equity book is inadvertently 18% overweight Technology. Which reference data management practice most directly prevents this class of classification inconsistency?
• Your quant team ingests daily satellite parking-lot counts for retail names. Which data infrastructure control is MOST critical before promoting the feed to production alpha?
• Entity resolution across news, filings, and market data for the same issuer fails during mergers. What is the BEST architectural response?
• An investment firm builds a feature store for alpha signals used by both research and live trading. What capability is essential?
• A sustainable investment fund manager is constructing an ESG scoring model for a portfolio of 200 global equities. She discovers that ESG ratings from MSCI, Sustainalytics, and Bloomberg ESG diverge by more than 40 points on the same company in some cases. What is the most technically sound approach to address this inter-rater disagreement in her data pipeline?
• A private equity firm's data engineering team is building a standardized performance reporting system across their portfolio of 85 private companies. They are encountering significant data collection friction. Which of the following represents the most fundamental structural challenge unique to private market data collection that cannot be solved by technology alone?
• An investment firm's technology team is designing an ingestion pipeline to process 10,000 earnings call transcripts, SEC filings, and analyst reports per month. They need to extract structured financial entities — revenue figures, guidance statements, and risk factors — and store them for downstream LLM querying and quantitative model consumption. Which pipeline architecture is most appropriate for this scale and use case?
• A global macro hedge fund wants to map the relationships between 5,000 corporate entities, their supply chain dependencies, executive board overlaps, regulatory filing cross-references, and counterparty exposures to identify systemic contagion risks. A data architect proposes replacing the current relational database with a knowledge graph. Which specific capability of knowledge graphs makes them architecturally superior to a relational database for this use case?
• A multi-asset investment firm receives daily data feeds from 12 different vendors covering equities, fixed income, and derivatives pricing. The Chief Data Officer wants to implement a formal data quality scoring framework to prioritize remediation efforts, drive vendor SLA negotiations, and demonstrate data governance maturity to regulators. Which foundational standard provides the most comprehensive and widely recognized framework for this purpose?
• A quantitative research team discovers that their backtesting results have been consistently inflated by look-ahead bias. Investigation reveals that their data warehouse stores only the most current version of fundamental data — when companies restate earnings, the restated figures overwrite the original reported values, making historical data appear as though the restated figures were known at the time of the original announcement. Which database design pattern specifically and completely solves this problem?
• Why must AI feature pipelines respect market data entitlements?
• A research tick store for microstructure features should optimize for:
• A global asset manager running equities, fixed income, commodities, and alternatives has centralized all market and portfolio data into a single data warehouse managed by a central data engineering team. Business users in each asset class report months-long delays for new data products, and data quality issues in one asset class frequently impact others. The CIO tasks the data architecture team with restructuring the data platform. Which architectural approach best addresses these organizational and technical pain points at enterprise scale?
• A proprietary trading firm needs to propagate trade execution events — occurring at sub-millisecond frequency during market hours — to downstream systems including a real-time P&L calculator, a risk aggregator, a regulatory reporting engine, and an ML-based market impact estimator. The current batch-based ETL pipeline introduces 15-minute latency that is incompatible with intraday risk limits. Which architectural pattern most effectively decouples event producers from consumers while supporting the required throughput and latency profile?
• A quantitative research team needs to store and query 20 years of tick-by-tick price data across 5,000 equity instruments, ingesting approximately 50 million records per trading day. Their primary access patterns are: (1) range scans over a specific symbol's price history for backtesting, (2) cross-sectional queries at a specific timestamp across all symbols, and (3) downsampling aggregations (OHLCV) at arbitrary frequencies. Which database technology is best suited for this workload?
• A mid-sized systematic macro fund is evaluating how to source real-time equity and FX market data for its algorithmic trading strategies. The infrastructure team is debating between subscribing to a commercial financial data API (e.g., Bloomberg B-PIPE, Refinitiv Elektron) and establishing direct exchange data feeds through a co-location provider. Both options have dramatically different cost, latency, and operational profiles. Which factors should primarily drive the architecture decision between these two approaches?
• Streaming feature compute for intraday signals requires:
• Data quality scorecards for investment AI should track:
• Knowledge-graph features for supply-chain risk are valuable because:
• A lakehouse pattern helps investment research AI by:
• Price and fundamental features for ML must adjust for corporate actions because:
• Alternative data contracts often include expiry/usage windows. Systems must:
• AI pipelines joining news to prices fail most often due to:
• OCR + layout models for scanned financial statements require:
• When collaborating across legal entities on data science, consider:
• Feature freshness SLAs should be set by:
• A data catalog helps quants by:
• Streaming corporate actions into research systems requires:
• Critical pricing inputs for AI risk models should:
• Lineage from raw filing → extracted table → feature → model should be:
• Time travel on research lake tables enables:
• Research/prod feature parity tests should:
• Real-time feature monitoring should alert on: