Interview Q&A

Short answer: A filter is a component that allows logic to be executed before or after parts of the request pipeline, such as authorization, action execution, or result processing.

Real-world example (ShopNest)

A ShopNest ValidateModelAttribute runs before actions and returns 400 if ModelState is invalid—same rule for every controller.

Say this in the interview

1. Define — one clear sentence (the short answer above).
2. Example — relate it to a project like ShopNest or your real work.
3. Trade-off — when you would not use it.

Short answer: Filters are ASP.NET Core’s plug-in points that allow you to inject logic before or after specific pipeline stages such as: Authorization Resource execution Action execution Results processing Exception handling They help you implement cross-cutting concerns without cluttering controllers or actions.

Explain a bit more

In enterprise systems, filters are indispensable for: Logging Validation Authorization Error handling Performance profiling Policy enforcement

Real-world example (ShopNest)

A ShopNest ValidateModelAttribute runs before actions and returns 400 if ModelState is invalid—same rule for every controller.

Say this in the interview

1. Define — one clear sentence (the short answer above).
2. Example — relate it to a project like ShopNest or your real work.
3. Trade-off — when you would not use it.

Short answer: Middleware is a component in the HTTP request pipeline that can: Handle requests, Pass requests to the next middleware, Or short-circuit the pipeline. Middleware can: Perform actions before and/or after the next middleware executes. Be used for logging, authentication, error handling, etc. Middleware executes in the order it's added in Program.cs.

Real-world example (ShopNest)

ShopNest pipeline order matters: exception handling → HTTPS → auth → authorization → endpoints. Auth must run before protected APIs.

Say this in the interview

1. Define — one clear sentence (the short answer above).
2. Example — relate it to a project like ShopNest or your real work.
3. Trade-off — when you would not use it.

Short answer: Historically, developers would sprinkle try-catch blocks, logging frameworks, and authorization checks directly inside controllers. This led to: Repetition Hard-to-test classes Cross-cutting concerns polluting business logic Fragile code Filters solve these by offering centralized, consistent, and reusable implementation points.

Real-world example (ShopNest)

A ShopNest ValidateModelAttribute runs before actions and returns 400 if ModelState is invalid—same rule for every controller.

Say this in the interview

1. Define — one clear sentence (the short answer above).
2. Example — relate it to a project like ShopNest or your real work.
3. Trade-off — when you would not use it.

Short answer: uthorization, Resource, Action, Exception, and Result filters.

Real-world example (ShopNest)

A ShopNest ValidateModelAttribute runs before actions and returns 400 if ModelState is invalid—same rule for every controller.

Say this in the interview

1. Define — one clear sentence (the short answer above).
2. Example — relate it to a project like ShopNest or your real work.
3. Trade-off — when you would not use it.

Short answer: uthorization checks directly inside controllers. This led to: Repetition Hard-to-test classes Cross-cutting concerns polluting business logic Fragile code Filters solve these by offering centralized, consistent, and reusable implementation points.

Real-world example (ShopNest)

A ShopNest ValidateModelAttribute runs before actions and returns 400 if ModelState is invalid—same rule for every controller.

Say this in the interview

1. Define — one clear sentence (the short answer above).
2. Example — relate it to a project like ShopNest or your real work.
3. Trade-off — when you would not use it.

Short answer: pp.UseMiddleware<YourMiddleware>(); pp.UseRouting(); pp.UseEndpoints(endpoints => { endpoints.MapControllers(); }); pp.Run(); In older versions (e.g., .NET Core 3.1), Startup.cs is used: public void Configure(IApplicationBuilder app, IWebHostEnvironment env) { pp.UseMiddleware<YourMiddleware>(); pp.UseRouting(); pp.UseEndpoints(endpoints => {………… pp.UseMiddleware<YourMiddleware>(); pp.UseRouting();…

Explain a bit more

pp.UseEndpoints(endpoints => { endpoints.MapControllers(); }); pp.Run(); In older versions (e.g., .NET Core 3.1), Startup.cs is used: public void Configure(IApplicationBuilder app, IWebHostEnvironment env) { pp.UseMiddleware<YourMiddleware>(); pp.UseRouting(); pp.UseEndpoints(endpoints => { endpoints.MapControllers(); }); } pp.UseMiddleware<YourMiddleware>(); pp.UseRouting(); pp.UseEndpoints(endpoints => { endpoints.MapControllers(); }); pp.Run(); In older versions (e.g., .NET Core 3.1), Startup.cs is used: public void Configure(IApplicationBuilder app, IWebHostEnvironment env) { pp.UseMiddleware<YourMiddleware>(); pp.UseRouting();…

Say this in the interview

1. Define — one clear sentence (the short answer above).
2. Example — relate it to a project like ShopNest or your real work.
3. Trade-off — when you would not use it.

Short answer: var app = builder.Build(); app.UseMiddleware<YourMiddleware>(); app.UseRouting(); app.UseEndpoints(endpoints => { endpoints.MapControllers(); }); app.Run(); In older versions (e.g., .NET Core 3.1), Startup.cs is used: public void Configure(IApplicationBuilder app, IWebHostEnvironment env) { app.UseMiddleware<YourMiddleware>(); app.UseRouting(); app.UseEndpoints(endpoints => { endpoints.MapControllers(); }); }

Example code

In ASP.NET Core 6+ (minimal hosting model), middleware is added in Program.cs: var builder = WebApplication.CreateBuilder(args);

Say this in the interview

1. Define — one clear sentence (the short answer above).
2. Example — relate it to a project like ShopNest or your real work.
3. Trade-off — when you would not use it.

Short answer: Filters promote reusability, separation of concerns, cleaner controllers, and easier maintenance.

Real-world example (ShopNest)

ShopNest registers AppDbContext as Scoped and IMemoryCache as Singleton. Putting DbContext in a Singleton causes threading bugs.

Say this in the interview

1. Define — one clear sentence (the short answer above).
2. Example — relate it to a project like ShopNest or your real work.
3. Trade-off — when you would not use it.

Short answer: Example – Using ServiceFilter [ServiceFilter(typeof(AuditLogFilter))] public class UserController : ControllerBase { }

Real-world example (ShopNest)

A ShopNest ValidateModelAttribute runs before actions and returns 400 if ModelState is invalid—same rule for every controller.

Say this in the interview

1. Define — one clear sentence (the short answer above).
2. Example — relate it to a project like ShopNest or your real work.
3. Trade-off — when you would not use it.

Short answer: uthorization Before everything Identity, JWT, RBAC Resource Before model binding Caching, request trimming ction Before/after action Logging, validation Exception On unhandled errors Global error handling Result Before/after result Response wrapping, formatting Endpoint (ASP.NET Core 7+) round minimal API endpoints Logging, validation 3.1… Authorization……… Filters These execute first, determining whether the user…

Explain a bit more

can access the route. public AuditLogFilter(ILogger<AuditLogFilter> logger)

Example code

{
_logger = logger;
}
public void OnActionExecuting(ActionExecutingContext context)
{ _logger.LogInformation("Request started at: {time}", DateTime.UtcNow); }
public void OnActionExecuted(ActionExecutedContext context)
{ _logger.LogInformation("Request ended at: {time}", DateTime.UtcNow); }
} 3.4 Exception Filters Centralizing error handling is vital in fintech and healthcare APIs. Example use cases: Returning consistent error envelopes Logging unhandled exceptions Protecting internal stack traces 3.5 Result Filters These wrap the response. Use cases: Response compression Standardizing API responses Masking sensitive data 3.6 Endpoint Filters (ASP.NET Core 7+) Primarily for minimal APIs. Example use cases: Request validation Logging Caching Response transformation pp.MapGet("/users/{id}", (int id) => GetUser(id)) .AddEndpointFilter(new LoggingEndpointFilter());

Real-world example (ShopNest)

A ShopNest ValidateModelAttribute runs before actions and returns 400 if ModelState is invalid—same rule for every controller.

Say this in the interview

1. Define — one clear sentence (the short answer above).
2. Example — relate it to a project like ShopNest or your real work.
3. Trade-off — when you would not use it.

Short answer: Authorization Filters Resource Filters Action Filters Exception Filters Result Filters Endpoint Filters (ASP.NET Core 7+)

Real-world example (ShopNest)

A ShopNest ValidateModelAttribute runs before actions and returns 400 if ModelState is invalid—same rule for every controller.

Say this in the interview

1. Define — one clear sentence (the short answer above).
2. Example — relate it to a project like ShopNest or your real work.
3. Trade-off — when you would not use it.

Short answer: Action filters wrap the action method; result filters wrap the returned result (like ObjectResult or ViewResult).

Example code

Action filters wrap the action method; result filters wrap the returned result (like ObjectResult or ViewResult).

Real-world example (ShopNest)

ShopNest registers AppDbContext as Scoped and IMemoryCache as Singleton. Putting DbContext in a Singleton causes threading bugs.

Say this in the interview

1. Define — one clear sentence (the short answer above).
2. Example — relate it to a project like ShopNest or your real work.
3. Trade-off — when you would not use it.

Short answer: Diagram) -------------------------- | Authorization Filter | -------------------------- -------------------------- | Resource Filter | -------------------------- -------------------------- | Model Binding Happens | -------------------------- -------------------------- | Action Filter | -------------------------- -------------------------- | Action Method | -------------------------- -------------------------- |…

Explain a bit more

Result Filter | -------------------------- -------------------------- | Response Returned | --------------------------

Real-world example (ShopNest)

ShopNest pipeline order matters: exception handling → HTTPS → auth → authorization → endpoints. Auth must run before protected APIs.

Say this in the interview

1. Define — one clear sentence (the short answer above).
2. Example — relate it to a project like ShopNest or your real work.
3. Trade-off — when you would not use it.

Short answer: wait _next(context); // Post-processing logic } } Register it: pp.UseMiddleware<MyCustomMiddleware>(); wait _next(context); // Post-processing logic } } Register it: pp.UseMiddleware<MyCustomMiddleware>(); wait _next(context); // Post-processing logic } } Register it: pp.UseMiddleware<MyCustomMiddleware>(); wait _next(context); // Post-processing logic } }… Register it: pp.UseMiddleware<MyCustomMiddleware>();

Real-world example (ShopNest)

ShopNest pipeline order matters: exception handling → HTTPS → auth → authorization → endpoints. Auth must run before protected APIs.

Say this in the interview

1. Define — one clear sentence (the short answer above).
2. Example — relate it to a project like ShopNest or your real work.
3. Trade-off — when you would not use it.

Short answer: Create a class with: A constructor accepting RequestDelegate An Invoke or InvokeAsync method public class MyCustomMiddleware { private readonly RequestDelegate _next; public MyCustomMiddleware(RequestDelegate next) => _next = next; public async Task InvokeAsync(HttpContext context) { // Pre-processing logic await _next(context); // Post-processing logic } } Register it: app.UseMiddleware<MyCustomMiddleware>();

Example code

Create a class with: A constructor accepting RequestDelegate An Invoke or InvokeAsync method public class MyCustomMiddleware
{
private readonly RequestDelegate _next;
public MyCustomMiddleware(RequestDelegate next) => _next = next; public async Task InvokeAsync(HttpContext context)
{ // Pre-processing logic await _next(context); // Post-processing logic }
} Register it: app.UseMiddleware<MyCustomMiddleware>();

Real-world example (ShopNest)

ShopNest pipeline order matters: exception handling → HTTPS → auth → authorization → endpoints. Auth must run before protected APIs.

Say this in the interview

1. Define — one clear sentence (the short answer above).
2. Example — relate it to a project like ShopNest or your real work.
3. Trade-off — when you would not use it.

Short answer: uthorization filters always run first. Intermediate Level

Real-world example (ShopNest)

A ShopNest ValidateModelAttribute runs before actions and returns 400 if ModelState is invalid—same rule for every controller.

Say this in the interview

1. Define — one clear sentence (the short answer above).
2. Example — relate it to a project like ShopNest or your real work.
3. Trade-off — when you would not use it.

Short answer: Building an AI career in 2026 requires strong fundamentals plus deployable projects. Learn core ML concepts, LLM workflows, and production practices such as evaluation and monitoring. Employers prioritize practical execution and portfolio depth over theory alone.

Step-by-step approach

1. Learn foundational Python, statistics, ML basics, and data handling workflows.
2. Master GenAI stack: prompts, embeddings, vector search, RAG, and evaluation.
3. Build and deploy projects with APIs, orchestration, guardrails, and monitoring.
4. Practice interview prep across coding, ML concepts, and AI system design.
5. Maintain a weekly learning loop with experiments, benchmarks, and release updates.

Real-world example

Neha was working at Flipkart and needed to handle this situation: how to become an ai agent developer. She prepared a clear plan with timelines, ownership, and expected outcomes before speaking to HR and her manager. Arjun, who had recently moved to Zoho, reviewed her approach and helped her tighten the messaging with measurable results. Within a few weeks, Neha achieved a better career outcome while preserving strong professional relationships.

Mistakes to avoid

• Acting without understanding policy, market context, or role expectations.
• Using generic claims instead of measurable evidence and concrete examples.
• Delaying communication and creating last-minute pressure for stakeholders.
• Relying only on certificates without publishing deployable, evaluated AI projects.

Toolliyo resources

• AI Engineer Roadmap
• Generative AI Engineer Roadmap
• Interview Q&A Tutorials

Ship demo projects with evaluation metrics; real evidence beats certificate-only positioning.

Short answer: Building an AI career in 2026 requires strong fundamentals plus deployable projects. Learn core ML concepts, LLM workflows, and production practices such as evaluation and monitoring. Employers prioritize practical execution and portfolio depth over theory alone.

Step-by-step approach

1. Learn foundational Python, statistics, ML basics, and data handling workflows.
2. Master GenAI stack: prompts, embeddings, vector search, RAG, and evaluation.
3. Build and deploy projects with APIs, orchestration, guardrails, and monitoring.
4. Practice interview prep across coding, ML concepts, and AI system design.
5. Maintain a weekly learning loop with experiments, benchmarks, and release updates.

Real-world example

Karan was working at Razorpay and needed to handle this situation: how to learn ai from scratch. She prepared a clear plan with timelines, ownership, and expected outcomes before speaking to HR and her manager. Isha, who had recently moved to PhonePe, reviewed her approach and helped her tighten the messaging with measurable results. Within a few weeks, Karan achieved a better career outcome while preserving strong professional relationships.

Mistakes to avoid

• Acting without understanding policy, market context, or role expectations.
• Using generic claims instead of measurable evidence and concrete examples.
• Delaying communication and creating last-minute pressure for stakeholders.
• Relying only on certificates without publishing deployable, evaluated AI projects.

Toolliyo resources

• AI Engineer Roadmap
• Generative AI Engineer Roadmap
• Interview Q&A Tutorials

Ship demo projects with evaluation metrics; real evidence beats certificate-only positioning.

Short answer: Building an AI career in 2026 requires strong fundamentals plus deployable projects. Learn core ML concepts, LLM workflows, and production practices such as evaluation and monitoring. Employers prioritize practical execution and portfolio depth over theory alone.

Step-by-step approach

1. Learn foundational Python, statistics, ML basics, and data handling workflows.
2. Master GenAI stack: prompts, embeddings, vector search, RAG, and evaluation.
3. Build and deploy projects with APIs, orchestration, guardrails, and monitoring.
4. Practice interview prep across coding, ML concepts, and AI system design.
5. Maintain a weekly learning loop with experiments, benchmarks, and release updates.

Real-world example

Meera was working at Freshworks and needed to handle this situation: how long does it take to become an ai engineer. She prepared a clear plan with timelines, ownership, and expected outcomes before speaking to HR and her manager. Rohit, who had recently moved to CRED, reviewed her approach and helped her tighten the messaging with measurable results. Within a few weeks, Meera achieved a better career outcome while preserving strong professional relationships.

Mistakes to avoid

• Acting without understanding policy, market context, or role expectations.
• Using generic claims instead of measurable evidence and concrete examples.
• Delaying communication and creating last-minute pressure for stakeholders.
• Relying only on certificates without publishing deployable, evaluated AI projects.

Toolliyo resources

• AI Engineer Roadmap
• Generative AI Engineer Roadmap
• Interview Q&A Tutorials

Ship demo projects with evaluation metrics; real evidence beats certificate-only positioning.

Short answer: Building an AI career in 2026 requires strong fundamentals plus deployable projects. Learn core ML concepts, LLM workflows, and production practices such as evaluation and monitoring. Employers prioritize practical execution and portfolio depth over theory alone.

Step-by-step approach

1. Learn foundational Python, statistics, ML basics, and data handling workflows.
2. Master GenAI stack: prompts, embeddings, vector search, RAG, and evaluation.
3. Build and deploy projects with APIs, orchestration, guardrails, and monitoring.
4. Practice interview prep across coding, ML concepts, and AI system design.
5. Maintain a weekly learning loop with experiments, benchmarks, and release updates.

Real-world example

Priya was working at Zoho and needed to handle this situation: what skills are required for ai jobs. She prepared a clear plan with timelines, ownership, and expected outcomes before speaking to HR and her manager. Rahul, who had recently moved to TCS, reviewed her approach and helped her tighten the messaging with measurable results. Within a few weeks, Priya achieved a better career outcome while preserving strong professional relationships.

Mistakes to avoid

• Acting without understanding policy, market context, or role expectations.
• Using generic claims instead of measurable evidence and concrete examples.
• Delaying communication and creating last-minute pressure for stakeholders.
• Relying only on certificates without publishing deployable, evaluated AI projects.

Toolliyo resources

• AI Engineer Roadmap
• Generative AI Engineer Roadmap
• Interview Q&A Tutorials

Ship demo projects with evaluation metrics; real evidence beats certificate-only positioning.

Short answer: Building an AI career in 2026 requires strong fundamentals plus deployable projects. Learn core ML concepts, LLM workflows, and production practices such as evaluation and monitoring. Employers prioritize practical execution and portfolio depth over theory alone.

Step-by-step approach

1. Learn foundational Python, statistics, ML basics, and data handling workflows.
2. Master GenAI stack: prompts, embeddings, vector search, RAG, and evaluation.
3. Build and deploy projects with APIs, orchestration, guardrails, and monitoring.
4. Practice interview prep across coding, ML concepts, and AI system design.
5. Maintain a weekly learning loop with experiments, benchmarks, and release updates.

Real-world example

Ananya was working at PhonePe and needed to handle this situation: best ai certifications. She prepared a clear plan with timelines, ownership, and expected outcomes before speaking to HR and her manager. Vikram, who had recently moved to Infosys, reviewed her approach and helped her tighten the messaging with measurable results. Within a few weeks, Ananya achieved a better career outcome while preserving strong professional relationships.

Mistakes to avoid

• Acting without understanding policy, market context, or role expectations.
• Using generic claims instead of measurable evidence and concrete examples.
• Delaying communication and creating last-minute pressure for stakeholders.
• Relying only on certificates without publishing deployable, evaluated AI projects.

Toolliyo resources

• AI Engineer Roadmap
• Generative AI Engineer Roadmap
• Interview Q&A Tutorials

Ship demo projects with evaluation metrics; real evidence beats certificate-only positioning.

Short answer: Salary negotiation works best when you combine market benchmarks with your business impact. Present a realistic range, explain your value with measurable outcomes, and stay collaborative with HR. This approach improves your chances of a better CTC without sounding rigid.

Step-by-step approach

1. Gather market salary data for your role, city, experience, and skill stack.
2. Document your strongest outcomes with numbers that prove business impact.
3. Set a target range and minimum acceptable figure before the discussion.
4. Present your ask confidently, then pause and let HR respond first.
5. If needed, negotiate components like bonus, variable pay, ESOPs, or review cycle.

Real-world example

Neha was working at CRED and needed to handle this situation: ai engineer salary in india. She prepared a clear plan with timelines, ownership, and expected outcomes before speaking to HR and her manager. Arjun, who had recently moved to Flipkart, reviewed her approach and helped her tighten the messaging with measurable results. Within a few weeks, Neha achieved a better career outcome while preserving strong professional relationships.

Numbers & benchmarks

• Typical switch hike in India often ranges from 25% to 60% based on demand and skill depth.
• Use a negotiation range width of around 10% to 15% instead of one rigid number.
• If possible, keep variable-heavy components below 20% for stable monthly cash flow.

Mistakes to avoid

• Acting without understanding policy, market context, or role expectations.
• Using generic claims instead of measurable evidence and concrete examples.
• Delaying communication and creating last-minute pressure for stakeholders.
• Accepting the first offer quickly without discussing structure, growth path, or review timeline.

Toolliyo resources

• AI Engineer Roadmap
• Generative AI Engineer Roadmap
• Interview Q&A Tutorials

Ship demo projects with evaluation metrics; real evidence beats certificate-only positioning.

Short answer: Building an AI career in 2026 requires strong fundamentals plus deployable projects. Learn core ML concepts, LLM workflows, and production practices such as evaluation and monitoring. Employers prioritize practical execution and portfolio depth over theory alone.

Step-by-step approach

1. Learn foundational Python, statistics, ML basics, and data handling workflows.
2. Master GenAI stack: prompts, embeddings, vector search, RAG, and evaluation.
3. Build and deploy projects with APIs, orchestration, guardrails, and monitoring.
4. Practice interview prep across coding, ML concepts, and AI system design.
5. Maintain a weekly learning loop with experiments, benchmarks, and release updates.

Real-world example

Karan was working at TCS and needed to handle this situation: ai engineer roadmap. She prepared a clear plan with timelines, ownership, and expected outcomes before speaking to HR and her manager. Isha, who had recently moved to Razorpay, reviewed her approach and helped her tighten the messaging with measurable results. Within a few weeks, Karan achieved a better career outcome while preserving strong professional relationships.

Mistakes to avoid

• Acting without understanding policy, market context, or role expectations.
• Using generic claims instead of measurable evidence and concrete examples.
• Delaying communication and creating last-minute pressure for stakeholders.
• Relying only on certificates without publishing deployable, evaluated AI projects.

Toolliyo resources

• AI Engineer Roadmap
• Generative AI Engineer Roadmap
• Interview Q&A Tutorials

Ship demo projects with evaluation metrics; real evidence beats certificate-only positioning.

Short answer: Building an AI career in 2026 requires strong fundamentals plus deployable projects. Learn core ML concepts, LLM workflows, and production practices such as evaluation and monitoring. Employers prioritize practical execution and portfolio depth over theory alone.

Step-by-step approach

1. Learn foundational Python, statistics, ML basics, and data handling workflows.
2. Master GenAI stack: prompts, embeddings, vector search, RAG, and evaluation.
3. Build and deploy projects with APIs, orchestration, guardrails, and monitoring.
4. Practice interview prep across coding, ML concepts, and AI system design.
5. Maintain a weekly learning loop with experiments, benchmarks, and release updates.

Real-world example

Meera was working at Infosys and needed to handle this situation: ai vs software engineering career. She prepared a clear plan with timelines, ownership, and expected outcomes before speaking to HR and her manager. Rohit, who had recently moved to Freshworks, reviewed her approach and helped her tighten the messaging with measurable results. Within a few weeks, Meera achieved a better career outcome while preserving strong professional relationships.

Mistakes to avoid

• Acting without understanding policy, market context, or role expectations.
• Using generic claims instead of measurable evidence and concrete examples.
• Delaying communication and creating last-minute pressure for stakeholders.
• Relying only on certificates without publishing deployable, evaluated AI projects.

Toolliyo resources

• AI Engineer Roadmap
• Generative AI Engineer Roadmap
• Interview Q&A Tutorials

Ship demo projects with evaluation metrics; real evidence beats certificate-only positioning.