diff --git a/decision_table_generator/src/decision_table_generator/core.py b/decision_table_generator/src/decision_table_generator/core.py
new file mode 100644
index 0000000..489bde9
--- /dev/null
+++ b/decision_table_generator/src/decision_table_generator/core.py
@@ -0,0 +1,107 @@
+from __future__ import annotations
+import logging
+from dataclasses import dataclass
+from typing import Any, Dict, List
+import pandas as pd
+
+
+logger = logging.getLogger(__name__)
+logger.setLevel(logging.INFO)
+
+
+class DecisionTableError(Exception):
+    """Custom exception for decision table generation errors."""
+
+
+@dataclass
+class DecisionConfig:
+    N_values: List[int]
+    warn_threshold: float
+    rerun_options: List[str]
+    unknown_handling: str
+
+    @classmethod
+    def from_dict(cls, data: Dict[str, Any]) -> DecisionConfig:
+        required_keys = {"N_values", "warn_threshold", "rerun_options", "unknown_handling"}
+        missing = required_keys - data.keys()
+        if missing:
+            raise ValueError(f"Missing fields in DecisionConfig: {missing}")
+
+        N_values = data.get("N_values")
+        warn_threshold = data.get("warn_threshold")
+        rerun_options = data.get("rerun_options")
+        unknown_handling = data.get("unknown_handling")
+
+        # Type validation
+        if not isinstance(N_values, list) or not all(isinstance(n, int) for n in N_values):
+            raise TypeError("N_values must be a list of integers.")
+        if not isinstance(warn_threshold, (float, int)):
+            raise TypeError("warn_threshold must be a float.")
+        if not isinstance(rerun_options, list) or not all(isinstance(r, str) for r in rerun_options):
+            raise TypeError("rerun_options must be a list of strings.")
+        if not isinstance(unknown_handling, str):
+            raise TypeError("unknown_handling must be a string.")
+
+        return cls(N_values=list(N_values),
+                   warn_threshold=float(warn_threshold),
+                   rerun_options=list(rerun_options),
+                   unknown_handling=str(unknown_handling))
+
+
+def generate_decision_table(config: Dict[str, Any]) -> List[Dict[str, Any]]:
+    """
+    Generate a structured decision table based on provided configuration.
+
+    Args:
+        config: A dictionary with keys matching DecisionConfig fields.
+
+    Returns:
+        A list of decision entries summarizing metrics per configuration.
+    """
+    logger.info("Starting generation of decision table.")
+    try:
+        cfg = DecisionConfig.from_dict(config)
+    except Exception as e:
+        logger.error(f"Invalid configuration: {e}")
+        raise DecisionTableError(str(e)) from e
+
+    # Construct decision rows (simulate with deterministic metrics for reproducibility)
+    rows: List[Dict[str, Any]] = []
+
+    for N in cfg.N_values:
+        for rerun_option in cfg.rerun_options:
+            warn_count = int(max(0, (cfg.warn_threshold / 100.0) * N))
+            rerun_helps = (N // 10) if rerun_option == 'on' else 0
+            rerun_shifts = (N // 20) if cfg.unknown_handling == 'separate_gate' else (N // 30)
+            rerun_hurts = max(0, (N // 40) - rerun_helps // 2)
+
+            row = {
+                'N': N,
+                'warn_threshold': cfg.warn_threshold,
+                'rerun_option': rerun_option,
+                'unknown_handling': cfg.unknown_handling,
+                'warn_count': warn_count,
+                'rerun_helps': rerun_helps,
+                'rerun_shifts': rerun_shifts,
+                'rerun_hurts': rerun_hurts,
+            }
+            rows.append(row)
+            logger.debug(f"Generated row: {row}")
+
+    df = pd.DataFrame(rows)
+    if df.empty:
+        logger.warning("Generated decision table is empty.")
+
+    # Ensure consistent column order
+    columns = [
+        'N', 'warn_threshold', 'rerun_option', 'unknown_handling',
+        'warn_count', 'rerun_helps', 'rerun_shifts', 'rerun_hurts'
+    ]
+    df = df[columns]
+
+    result = df.to_dict(orient='records')
+    # Simple CI-ready assertion: ensure fields exist
+    assert all(set(columns) == set(rec.keys()) for rec in result), "Mismatch in expected fields."
+
+    logger.info(f"Decision table generated with {len(result)} entries.")
+    return result