llm evals

app.py

@@ -1,12 +1,13 @@
 import gradio as gr
-from rag import rbc_product
-from tool import rival_product
-from graphrag import marketingPlan
-from knowledge import graph
-from pii import derisk
+
 from classify import judge
 from entity import resolve
+from graphrag import marketingPlan
 from human import email, feedback
+from knowledge import graph
+from pii import derisk
+from rag import rbc_product
+from tool import rival_product
 
 # Define the Google Analytics script
 head = """
@@ -94,11 +95,12 @@ Other Links:
 
     gr.Examples(
       [
-        ["Low APR and great customer service. I would highly recommend if you’re looking for a great credit card company and looking to rebuild your credit. I have had my credit limit increased annually and the annual fee is very low."]
+        [
+          "Low APR and great customer service. I would highly recommend if you’re looking for a great credit card company and looking to rebuild your credit. I have had my credit limit increased annually and the annual fee is very low."]
       ],
       [in_verbatim]
     )
-    btn_recommend=gr.Button("Recommend")
+    btn_recommend = gr.Button("Recommend")
     btn_recommend.click(fn=rival_product, inputs=in_verbatim, outputs=out_product)
 
     gr.Markdown("""
@@ -254,7 +256,7 @@ Representative: "Confirmed. Your next payment of $200 will process May 1st. A co
 
 Customer: "No, thank you."
          """
-          ]
+        ]
       ],
       [in_verbatim]
     )
@@ -262,7 +264,6 @@ Customer: "No, thank you."
     btn_clear = gr.ClearButton(components=[out_product])
     btn_recommend.click(fn=graph, inputs=[in_verbatim, out_product], outputs=out_product)
 
-
     gr.Markdown("""
 Example of Customer Profile in Graph
 =================
@@ -306,15 +307,15 @@ Once created, knowledge graphs can be repurposed across multiple use cases (e.g.
     gr.Examples(
       [
         [
-        """
-        He Hua (Hua Hua) Director
-        hehua@chengdu.com
-        +86-28-83505513
-
-        Alternative Address Format:
-        Xiongmao Ave West Section, Jinniu District (listed in some records as 610016 postcode)
-        """
-          ]
+          """
+          He Hua (Hua Hua) Director
+          hehua@chengdu.com
+          +86-28-83505513
+  
+          Alternative Address Format:
+          Xiongmao Ave West Section, Jinniu District (listed in some records as 610016 postcode)
+          """
+        ]
       ],
       [in_verbatim]
     )
@@ -333,7 +334,6 @@ Removes noise (e.g., irrelevant names or addresses) to make datasets cleaner and
 Allows downstream tasks (like sentiment analysis or topic modeling) to focus on content rather than personal identifiers.
     """)
 
-
   with gr.Tab("Segmentation"):
     gr.Markdown("""
     Objective: Streamline Customer Insights: Auto-Classify Feedback for Product Optimization
@@ -353,14 +353,14 @@ Allows downstream tasks (like sentiment analysis or topic modeling) to focus on
     gr.Examples(
       [
         [
-        """
-"The online portal makes managing my mortgage payments so convenient.";
-"RBC offer great mortgage for my home with competitive rate thank you";
-"Low interest rate compared to other cards I’ve used. Highly recommend for responsible spenders.";
-"The mobile check deposit feature saves me so much time. Banking made easy!";
-"Affordable premiums with great coverage. Switched from my old provider and saved!"
-        """
-          ]
+          """
+  "The online portal makes managing my mortgage payments so convenient.";
+  "RBC offer great mortgage for my home with competitive rate thank you";
+  "Low interest rate compared to other cards I’ve used. Highly recommend for responsible spenders.";
+  "The mobile check deposit feature saves me so much time. Banking made easy!";
+  "Affordable premiums with great coverage. Switched from my old provider and saved!"
+          """
+        ]
       ],
       [in_verbatim]
     )
@@ -444,7 +444,7 @@ Customer: "No, thank you."
       ],
       [in_verbatim]
     )
-    btn_recommend=gr.Button("Resolve")
+    btn_recommend = gr.Button("Resolve")
     btn_recommend.click(fn=resolve, inputs=in_verbatim, outputs=out_product)
 
     gr.Markdown("""
@@ -483,7 +483,9 @@ For example, Comcast reduced repeat service calls by 17% after deploying entity
 
     gr.Examples(
       [
-        ["""My mortgage was assumed by Bank of America when Countrywide mortgages ceased to do business. My mortgage increased without any explanation. When I inquired, they stumbled and gave me the run around. I’d NEVER do business with Bank of America again""", "MORT"],
+        [
+          """My mortgage was assumed by Bank of America when Countrywide mortgages ceased to do business. My mortgage increased without any explanation. When I inquired, they stumbled and gave me the run around. I’d NEVER do business with Bank of America again""",
+          "MORT"],
         ["my credit card limit is too low, I need a card with bigger limit and low fee", "CARD"]
       ],
       [in_verbatim, in_campaign]
@@ -541,50 +543,50 @@ For example, Comcast reduced repeat service calls by 17% after deploying entity
     btn_recommend.click(fn=rbc_product, inputs=in_verbatim, outputs=out_product)
 
     gr.Markdown("""
-Companies pour millions into product catalogs, marketing funnels, and user acquisition—yet many still face the same challenge:
-==================
-### 📉 Pain points:
-- High bounce rates and low conversion despite heavy traffic
-- Customers struggle to find relevant products on their own
-- One-size-fits-all promotions result in wasted ad spend and poor ROI
-
-### 🧩 The real question:
-What if your product catalog could *adapt itself* to each user in real time—just like your best salesperson would?
-
-### 🎯 The customer need:
-Businesses need a way to dynamically personalize product discovery, so every customer sees the most relevant items—without manually configuring hundreds of rules.
-
-## ✅ Enter: Product Recommender Systems
-
-By analyzing behavioral data, preferences, and historical purchases, a recommender engine surfaces what each user is most likely to want—boosting engagement and revenue.
-
-### 📌 Real-world use cases:
-- **Amazon** attributes up to 35% of its revenue to its recommender system, which tailors the home page, emails, and checkout cross-sells per user.
-- **Netflix** leverages personalized content recommendations to reduce churn and increase watch time—saving the company over $1B annually in retention value.
-- **Stitch Fix** uses machine learning-powered recommendations to curate clothing boxes tailored to individual style profiles—scaling personal styling.
-
-### 💡 Business benefits:
-- Higher conversion rates through relevant discovery
-- Increased average order value (AOV) via cross-sell and upsell
-- Improved retention and lower customer acquisition cost (CAC)
-
-If your product discovery experience isn’t working as hard as your marketing budget, it’s time to make your catalog intelligent—with recommendations that convert.
+    Companies pour millions into product catalogs, marketing funnels, and user acquisition—yet many still face the same challenge:
+    ==================
+    ### 📉 Pain points:
+    - High bounce rates and low conversion despite heavy traffic
+    - Customers struggle to find relevant products on their own
+    - One-size-fits-all promotions result in wasted ad spend and poor ROI
+    
+    ### 🧩 The real question:
+    What if your product catalog could *adapt itself* to each user in real time—just like your best salesperson would?
+    
+    ### 🎯 The customer need:
+    Businesses need a way to dynamically personalize product discovery, so every customer sees the most relevant items—without manually configuring hundreds of rules.
+    
+    ## ✅ Enter: Product Recommender Systems
+    
+    By analyzing behavioral data, preferences, and historical purchases, a recommender engine surfaces what each user is most likely to want—boosting engagement and revenue.
+    
+    ### 📌 Real-world use cases:
+    - **Amazon** attributes up to 35% of its revenue to its recommender system, which tailors the home page, emails, and checkout cross-sells per user.
+    - **Netflix** leverages personalized content recommendations to reduce churn and increase watch time—saving the company over $1B annually in retention value.
+    - **Stitch Fix** uses machine learning-powered recommendations to curate clothing boxes tailored to individual style profiles—scaling personal styling.
+    
+    ### 💡 Business benefits:
+    - Higher conversion rates through relevant discovery
+    - Increased average order value (AOV) via cross-sell and upsell
+    - Improved retention and lower customer acquisition cost (CAC)
+    
+    If your product discovery experience isn’t working as hard as your marketing budget, it’s time to make your catalog intelligent—with recommendations that convert.
     """)
 
-  with gr.Tab("LLM Evals"):
+  with gr.Tab("Eval"):
     gr.Markdown("""
-🏦 LLMs for Application Security in Personal Banking
-====================
-What happens when your generative AI exposes customer data before you even launch?
-
-LLM evals reduce security risks in generative AI banking apps by identifying vulnerabilities and guiding secure fixes.
-
-Personal banking apps increasingly rely on generative AI—but insecure logic and hallucinations expose sensitive customer data. LLM evals help assess code and AI-generated responses for correctness, task completion, hallucination risk, and safety—enabling proactive guardrails against vulnerabilities before deployment.
-
-I’ve led cross-functional model risk initiatives, building pipelines that transform LLM evaluations into automated alerts and remediation workflows—strengthening regulatory compliance and protecting customer trust.
-
-Using open-source frameworks, I identify flaws in LLM prompt and translate risks into explainable insights for business, risk, and engineering stakeholders.
-https://postimg.cc/3WtG4ZK2
+    🏦 LLMs for Application Security in Personal Banking
+    ====================
+    What happens when your generative AI exposes customer data before you even launch?
+    
+    LLM evals reduce security risks in generative AI banking apps by identifying vulnerabilities and guiding secure fixes.
+    
+    Personal banking apps increasingly rely on generative AI—but insecure logic and hallucinations expose sensitive customer data. LLM evals help assess code and AI-generated responses for correctness, task completion, hallucination risk, and safety—enabling proactive guardrails against vulnerabilities before deployment.
+    
+    I’ve led cross-functional model risk initiatives, building pipelines that transform LLM evaluations into automated alerts and remediation workflows—strengthening regulatory compliance and protecting customer trust.
+    
+    Using open-source frameworks, I identify flaws in LLM prompt and translate risks into explainable insights for business, risk, and engineering stakeholders.
+    https://postimg.cc/3WtG4ZK2
     """)
 
-demo.launch(allowed_paths=["."])
+demo.launch(allowed_paths=["."])