diff --git a/json_to_files.py b/json_to_files.py
new file mode 100644
index 00000000..e69de29b
diff --git a/questions/140_bernoulli-naive-bayes-classifier/description.md b/questions/140_bernoulli-naive-bayes-classifier/description.md
new file mode 100644
index 00000000..bcd62b6c
--- /dev/null
+++ b/questions/140_bernoulli-naive-bayes-classifier/description.md
@@ -0,0 +1 @@
+Write a Python class to implement the Bernoulli Naive Bayes classifier for binary (0/1) feature data. Your class should have two methods: `forward(self, X, y)` to train on the input data (X: 2D NumPy array of binary features, y: 1D NumPy array of class labels) and `predict(self, X)` to output predicted labels for a 2D test matrix X. Use Laplace smoothing (parameter: smoothing=1.0). Return predictions as a NumPy array. Only use NumPy. Predictions must be binary (0 or 1) and you must handle cases where the training data contains only one class. All log/likelihood calculations should use log probabilities for numerical stability.
diff --git a/questions/140_bernoulli-naive-bayes-classifier/example.json b/questions/140_bernoulli-naive-bayes-classifier/example.json
new file mode 100644
index 00000000..fc8cc328
--- /dev/null
+++ b/questions/140_bernoulli-naive-bayes-classifier/example.json
@@ -0,0 +1,5 @@
+{
+  "input": "X = np.array([[1, 0, 1], [1, 1, 0], [0, 0, 1], [0, 1, 0], [1, 1, 1]]); y = np.array([1, 1, 0, 0, 1])\nmodel = NaiveBayes(smoothing=1.0)\nmodel.forward(X, y)\nprint(model.predict(np.array([[1, 0, 1]])))",
+  "output": "[1]",
+  "reasoning": "The model learns class priors and feature probabilities with Laplace smoothing. For [1, 0, 1], the posterior for class 1 is higher, so the model predicts 1."
+}
diff --git a/questions/140_bernoulli-naive-bayes-classifier/learn.md b/questions/140_bernoulli-naive-bayes-classifier/learn.md
new file mode 100644
index 00000000..f6e05ccc
--- /dev/null
+++ b/questions/140_bernoulli-naive-bayes-classifier/learn.md
@@ -0,0 +1,51 @@
+# **Naive Bayes Classifier**
+
+## **1. Definition**
+
+Naive Bayes is a **probabilistic machine learning algorithm** used for **classification tasks**. It is based on **Bayes' Theorem**, which describes the probability of an event based on prior knowledge of related events.
+
+The algorithm assumes that:
+- **Features are conditionally independent** given the class label (the "naive" assumption).
+- It calculates the posterior probability for each class and assigns the class with the **highest posterior** to the sample.
+
+---
+
+## **2. Bayes' Theorem**
+
+Bayes' Theorem is given by:
+
+$$
+P(C | X) = \frac{P(X | C) \times P(C)}{P(X)}
+$$
+
+Where:
+- $P(C | X)$ **Posterior** probability: the probability of class $C $ given the feature vector $X$
+- $P(X | C)$ → **Likelihood**: the probability of the data $X$ given the class
+- $P(C)$ → **Prior** probability: the initial probability of class $C$ before observing any data
+- $ P(X)$ → **Evidence**: the total probability of the data across all classes (acts as a normalizing constant)
+
+Since $P(X)$ is the same for all classes during comparison, it can be ignored, simplifying the formula to:
+
+$$
+P(C | X) \propto P(X | C) \times P(C)
+$$
+---
+
+### 3 **Bernoulli Naive Bayes**
+- Used for **binary data** (features take only 0 or 1 values).
+- The likelihood is given by:
+
+$$
+P(X | C) = \prod_{i=1}^{n} P(x_i | C)^{x_i} \cdot (1 - P(x_i | C))^{1 - x_i}
+$$
+
+---
+
+## **4. Applications of Naive Bayes**
+
+- **Text Classification:** Spam detection, sentiment analysis, and news categorization.
+- **Document Categorization:** Sorting documents by topic.
+- **Fraud Detection:** Identifying fraudulent transactions or behaviors.
+- **Recommender Systems:** Classifying users into preference groups.
+
+--- 
diff --git a/questions/140_bernoulli-naive-bayes-classifier/meta.json b/questions/140_bernoulli-naive-bayes-classifier/meta.json
new file mode 100644
index 00000000..2a87586d
--- /dev/null
+++ b/questions/140_bernoulli-naive-bayes-classifier/meta.json
@@ -0,0 +1,15 @@
+{
+  "id": "140",
+  "title": "Bernoulli Naive Bayes Classifier",
+  "difficulty": "medium",
+  "category": "Machine Learning",
+  "video": "",
+  "likes": "0",
+  "dislikes": "0",
+  "contributor": [
+    {
+      "profile_link": "https://github.com/Coder1010ayush",
+      "name": "Coder1010ayush"
+    }
+  ]
+}
diff --git a/questions/140_bernoulli-naive-bayes-classifier/solution.py b/questions/140_bernoulli-naive-bayes-classifier/solution.py
new file mode 100644
index 00000000..2c99379d
--- /dev/null
+++ b/questions/140_bernoulli-naive-bayes-classifier/solution.py
@@ -0,0 +1,30 @@
+import numpy as np
+
+class NaiveBayes():
+    def __init__(self, smoothing=1.0):
+        self.smoothing = smoothing
+        self.classes = None
+        self.priors = None
+        self.likelihoods = None
+
+    def forward(self, X, y):
+        self.classes, class_counts = np.unique(y, return_counts=True)
+        self.priors = {cls: np.log(class_counts[i] / len(y)) for i, cls in enumerate(self.classes)}
+        self.likelihoods = {}
+        for cls in self.classes:
+            X_cls = X[y == cls]
+            prob = (np.sum(X_cls, axis=0) + self.smoothing) / (X_cls.shape[0] + 2 * self.smoothing)
+            self.likelihoods[cls] = (np.log(prob), np.log(1 - prob))
+
+    def _compute_posterior(self, sample):
+        posteriors = {}
+        for cls in self.classes:
+            posterior = self.priors[cls]
+            prob_1, prob_0 = self.likelihoods[cls]
+            likelihood = np.sum(sample * prob_1 + (1 - sample) * prob_0)
+            posterior += likelihood
+            posteriors[cls] = posterior
+        return max(posteriors, key=posteriors.get)
+
+    def predict(self, X):
+        return np.array([self._compute_posterior(sample) for sample in X])
diff --git a/questions/140_bernoulli-naive-bayes-classifier/starter_code.py b/questions/140_bernoulli-naive-bayes-classifier/starter_code.py
new file mode 100644
index 00000000..09c9820e
--- /dev/null
+++ b/questions/140_bernoulli-naive-bayes-classifier/starter_code.py
@@ -0,0 +1,14 @@
+import numpy as np
+
+class NaiveBayes():
+    def __init__(self, smoothing=1.0):
+        # Initialize smoothing
+        pass
+
+    def forward(self, X, y):
+        # Fit model to binary features X and labels y
+        pass
+
+    def predict(self, X):
+        # Predict class labels for test set X
+        pass
diff --git a/questions/140_bernoulli-naive-bayes-classifier/tests.json b/questions/140_bernoulli-naive-bayes-classifier/tests.json
new file mode 100644
index 00000000..e74c2f58
--- /dev/null
+++ b/questions/140_bernoulli-naive-bayes-classifier/tests.json
@@ -0,0 +1,22 @@
+[
+  {
+    "test": "import numpy as np\nmodel = NaiveBayes(smoothing=1.0)\nX = np.array([[1, 0, 1], [1, 1, 0], [0, 0, 1], [0, 1, 0], [1, 1, 1]])\ny = np.array([1, 1, 0, 0, 1])\nmodel.forward(X, y)\nprint(model.predict(np.array([[1, 0, 1]])))",
+    "expected_output": "[1]"
+  },
+  {
+    "test": "import numpy as np\nmodel = NaiveBayes(smoothing=1.0)\nX = np.array([[0], [1], [0], [1]])\ny = np.array([0, 1, 0, 1])\nmodel.forward(X, y)\nprint(model.predict(np.array([[0], [1]])))",
+    "expected_output": "[0 1]"
+  },
+  {
+    "test": "import numpy as np\nmodel = NaiveBayes(smoothing=1.0)\nX = np.array([[0, 0], [1, 0], [0, 1]])\ny = np.array([0, 1, 0])\nmodel.forward(X, y)\nprint(model.predict(np.array([[1, 1]])))",
+    "expected_output": "[0]"
+  },
+  {
+    "test": "import numpy as np\nnp.random.seed(42)\nmodel = NaiveBayes(smoothing=1.0)\nX = np.random.randint(0, 2, (100, 5))\ny = np.random.choice([0, 1], size=100)\nmodel.forward(X, y)\nX_test = np.random.randint(0, 2, (10, 5))\npred = model.predict(X_test)\nprint(pred.shape)",
+    "expected_output": "(10,)"
+  },
+  {
+    "test": "import numpy as np\nmodel = NaiveBayes(smoothing=1.0)\nX = np.random.randint(0, 2, (10, 3))\ny = np.zeros(10)\nmodel.forward(X, y)\nX_test = np.random.randint(0, 2, (3, 3))\nprint(model.predict(X_test))",
+    "expected_output": "[0, 0, 0]"
+  }
+]
diff --git a/utils/convert_single_question.py b/utils/convert_single_question.py
new file mode 100644
index 00000000..ef9f9fae
--- /dev/null
+++ b/utils/convert_single_question.py
@@ -0,0 +1,209 @@
+#!/usr/bin/env python
+"""
+convert_single_question.py
+──────────────────────────
+Paste ONE question dict into QUESTION_DICT and run:
+
+    python utils/convert_single_question.py
+
+The script splits it into:
+
+questions/<id>_<slug>/
+    ├─ meta.json
+    ├─ description.md
+    ├─ learn.md
+    ├─ starter_code.py
+    ├─ solution.py
+    ├─ example.json
+    ├─ tests.json
+    ├─ tinygrad/   (optional)
+    └─ pytorch/    (optional)
+"""
+
+import base64
+import json
+import pathlib
+import re
+from typing import Any, Dict
+
+# ── 1️⃣  EDIT YOUR QUESTION HERE ────────────────────────────────────────────
+QUESTION_DICT: Dict[str, Any] = {
+    'id':'140',
+    "description": "Write a Python class to implement the Bernoulli Naive Bayes classifier for binary (0/1) feature data. Your class should have two methods: `forward(self, X, y)` to train on the input data (X: 2D NumPy array of binary features, y: 1D NumPy array of class labels) and `predict(self, X)` to output predicted labels for a 2D test matrix X. Use Laplace smoothing (parameter: smoothing=1.0). Return predictions as a NumPy array. Only use NumPy. Predictions must be binary (0 or 1) and you must handle cases where the training data contains only one class. All log/likelihood calculations should use log probabilities for numerical stability.",
+    "test_cases": [
+        {
+            "test": "import numpy as np\nmodel = NaiveBayes(smoothing=1.0)\nX = np.array([[1, 0, 1], [1, 1, 0], [0, 0, 1], [0, 1, 0], [1, 1, 1]])\ny = np.array([1, 1, 0, 0, 1])\nmodel.forward(X, y)\nprint(model.predict(np.array([[1, 0, 1]])))",
+            "expected_output": "[1]"
+        },
+        {
+            "test": "import numpy as np\nmodel = NaiveBayes(smoothing=1.0)\nX = np.array([[0], [1], [0], [1]])\ny = np.array([0, 1, 0, 1])\nmodel.forward(X, y)\nprint(model.predict(np.array([[0], [1]])))",
+            "expected_output": "[0 1]"
+        },
+        {
+            "test": "import numpy as np\nmodel = NaiveBayes(smoothing=1.0)\nX = np.array([[0, 0], [1, 0], [0, 1]])\ny = np.array([0, 1, 0])\nmodel.forward(X, y)\nprint(model.predict(np.array([[1, 1]])))",
+            "expected_output": "[0]"
+        },
+        {
+            "test": "import numpy as np\nnp.random.seed(42)\nmodel = NaiveBayes(smoothing=1.0)\nX = np.random.randint(0, 2, (100, 5))\ny = np.random.choice([0, 1], size=100)\nmodel.forward(X, y)\nX_test = np.random.randint(0, 2, (10, 5))\npred = model.predict(X_test)\nprint(pred.shape)",
+            "expected_output": "(10,)"
+        },
+        {
+            "test": "import numpy as np\nmodel = NaiveBayes(smoothing=1.0)\nX = np.random.randint(0, 2, (10, 3))\ny = np.zeros(10)\nmodel.forward(X, y)\nX_test = np.random.randint(0, 2, (3, 3))\nprint(model.predict(X_test))",
+            "expected_output": "[0, 0, 0]"
+        }
+    ],
+    "solution": "import numpy as np\n\nclass NaiveBayes():\n    def __init__(self, smoothing=1.0):\n        self.smoothing = smoothing\n        self.classes = None\n        self.priors = None\n        self.likelihoods = None\n\n    def forward(self, X, y):\n        self.classes, class_counts = np.unique(y, return_counts=True)\n        self.priors = {cls: np.log(class_counts[i] / len(y)) for i, cls in enumerate(self.classes)}\n        self.likelihoods = {}\n        for cls in self.classes:\n            X_cls = X[y == cls]\n            prob = (np.sum(X_cls, axis=0) + self.smoothing) / (X_cls.shape[0] + 2 * self.smoothing)\n            self.likelihoods[cls] = (np.log(prob), np.log(1 - prob))\n\n    def _compute_posterior(self, sample):\n        posteriors = {}\n        for cls in self.classes:\n            posterior = self.priors[cls]\n            prob_1, prob_0 = self.likelihoods[cls]\n            likelihood = np.sum(sample * prob_1 + (1 - sample) * prob_0)\n            posterior += likelihood\n            posteriors[cls] = posterior\n        return max(posteriors, key=posteriors.get)\n\n    def predict(self, X):\n        return np.array([self._compute_posterior(sample) for sample in X])",
+    "example": {
+        "input": "X = np.array([[1, 0, 1], [1, 1, 0], [0, 0, 1], [0, 1, 0], [1, 1, 1]]); y = np.array([1, 1, 0, 0, 1])\nmodel = NaiveBayes(smoothing=1.0)\nmodel.forward(X, y)\nprint(model.predict(np.array([[1, 0, 1]])))",
+        "output": "[1]",
+        "reasoning": "The model learns class priors and feature probabilities with Laplace smoothing. For [1, 0, 1], the posterior for class 1 is higher, so the model predicts 1."
+    },
+    "category": "Machine Learning",
+    "starter_code": "import numpy as np\n\nclass NaiveBayes():\n    def __init__(self, smoothing=1.0):\n        # Initialize smoothing\n        pass\n\n    def forward(self, X, y):\n        # Fit model to binary features X and labels y\n        pass\n\n    def predict(self, X):\n        # Predict class labels for test set X\n        pass",
+    "title": "Bernoulli Naive Bayes Classifier",
+    "learn_section":r"""# **Naive Bayes Classifier**
+
+## **1. Definition**
+
+Naive Bayes is a **probabilistic machine learning algorithm** used for **classification tasks**. It is based on **Bayes' Theorem**, which describes the probability of an event based on prior knowledge of related events.
+
+The algorithm assumes that:
+- **Features are conditionally independent** given the class label (the "naive" assumption).
+- It calculates the posterior probability for each class and assigns the class with the **highest posterior** to the sample.
+
+---
+
+## **2. Bayes' Theorem**
+
+Bayes' Theorem is given by:
+
+$$
+P(C | X) = \frac{P(X | C) \times P(C)}{P(X)}
+$$
+
+Where:
+- $P(C | X)$ **Posterior** probability: the probability of class $C $ given the feature vector $X$
+- $P(X | C)$ → **Likelihood**: the probability of the data $X$ given the class
+- $P(C)$ → **Prior** probability: the initial probability of class $C$ before observing any data
+- $ P(X)$ → **Evidence**: the total probability of the data across all classes (acts as a normalizing constant)
+
+Since $P(X)$ is the same for all classes during comparison, it can be ignored, simplifying the formula to:
+
+$$
+P(C | X) \propto P(X | C) \times P(C)
+$$
+---
+
+### 3 **Bernoulli Naive Bayes**
+- Used for **binary data** (features take only 0 or 1 values).
+- The likelihood is given by:
+
+$$
+P(X | C) = \prod_{i=1}^{n} P(x_i | C)^{x_i} \cdot (1 - P(x_i | C))^{1 - x_i}
+$$
+
+---
+
+## **4. Applications of Naive Bayes**
+
+- **Text Classification:** Spam detection, sentiment analysis, and news categorization.
+- **Document Categorization:** Sorting documents by topic.
+- **Fraud Detection:** Identifying fraudulent transactions or behaviors.
+- **Recommender Systems:** Classifying users into preference groups.
+
+--- """,
+    "contributor": [
+        {
+            "profile_link": "https://github.com/moe18",
+            "name": "Moe Chabot"
+        }
+    ],
+    "likes": "0",
+    "dislikes": "0",
+    "difficulty": "medium",
+    "video":''
+}
+
+# ────────────────────────────────────────────────────────────────────────────
+
+
+# ---------- helpers ---------------------------------------------------------
+def slugify(text: str) -> str:
+    text = re.sub(r"[^0-9A-Za-z]+", "-", text.lower())
+    return re.sub(r"-{2,}", "-", text).strip("-")[:50]
+
+
+def maybe_b64(s: str) -> str:
+    try:
+        if len(s) % 4 == 0 and re.fullmatch(r"[0-9A-Za-z+/=\n\r]+", s):
+            return base64.b64decode(s).decode("utf-8")
+    except Exception:
+        pass
+    return s
+
+
+def write_text(path: pathlib.Path, content: str) -> None:
+    path.parent.mkdir(parents=True, exist_ok=True)
+    path.write_text(content.rstrip("\n") + "\n", encoding="utf-8")
+
+
+def write_json(path: pathlib.Path, obj: Any) -> None:
+    write_text(path, json.dumps(obj, indent=2, ensure_ascii=False))
+
+
+# ---------- converter -------------------------------------------------------
+def convert_one(q: Dict[str, Any]) -> None:
+    folder = pathlib.Path("questions") / f"{q['id']}_{slugify(q['title'])}"
+    folder.mkdir(parents=True, exist_ok=True)
+
+    # meta.json
+    meta = {
+        "id": q["id"],
+        "title": q["title"],
+        "difficulty": q["difficulty"],
+        "category": q["category"],
+        "video": q.get("video", ""),
+        "likes": q.get("likes", "0"),
+        "dislikes": q.get("dislikes", "0"),
+        "contributor": q.get("contributor", []),
+    }
+    for opt in ("tinygrad_difficulty", "pytorch_difficulty", "marimo_link"):
+        if opt in q:
+            meta[opt] = q[opt]
+    write_json(folder / "meta.json", meta)
+
+    # core files
+    write_text(folder / "description.md", q["description"])
+    write_text(folder / "learn.md", q["learn_section"])
+    write_text(folder / "starter_code.py", q["starter_code"])
+    write_text(folder / "solution.py", q["solution"])
+    write_json(folder / "example.json", q["example"])
+    write_json(folder / "tests.json", q["test_cases"])
+
+    # optional language-specific extras
+    for lang in ("tinygrad", "pytorch"):
+        sc, so, tc = (f"{lang}_starter_code", f"{lang}_solution", f"{lang}_test_cases")
+        if any(k in q for k in (sc, so, tc)):
+            sub = folder / lang
+            if sc in q:
+                write_text(sub / "starter_code.py", maybe_b64(q[sc]))
+            if so in q:
+                write_text(sub / "solution.py", maybe_b64(q[so]))
+            if tc in q:
+                write_json(sub / "tests.json", q[tc])
+
+    # success message (relative if possible)
+    try:
+        rel = folder.relative_to(pathlib.Path.cwd())
+    except ValueError:
+        rel = folder
+    print(f"✓  Created {rel}")
+
+
+# ---------- main ------------------------------------------------------------
+def main() -> None:
+    convert_one(QUESTION_DICT)
+
+
+if __name__ == "__main__":
+    main()