senior-backend

name	senior-backend
description	Comprehensive backend development skill for building scalable API systems using Node.js, Next.js API routes, tRPC, PostgreSQL, Prisma, Redis. Use this skill whenever designing APIs, writing database queries, creating server-side logic, handling authentication, building tRPC routers, optimizing database performance, or any backend/server task.

Senior Backend

Core Principles

Build secure, scalable, maintainable backend systems. Every API must be: typed end-to-end, secured by default, observable, and resilient.

Tech Stack (This Project)

• Runtime: Node.js (via Next.js serverless)
• API Layer: tRPC (type-safe) + Next.js API routes (webhooks, health)
• Database: PostgreSQL via Supabase
• ORM: Prisma
• Cache: Upstash Redis
• Queue: Inngest (background jobs)
• Auth: NextAuth.js v5

tRPC Router Patterns

Router Structure (one file per domain)

// server/routers/session.router.ts
import { z } from 'zod';
import { router, protectedProcedure, publicProcedure } from '../trpc';

const createSessionSchema = z.object({
  type: z.enum(['intake', 'standard', 'crisis']),
  scheduledAt: z.string().datetime().optional(),
});

export const sessionRouter = router({
  // Queries — read data
  list: protectedProcedure
    .input(z.object({
      cursor: z.string().optional(),
      limit: z.number().min(1).max(50).default(20),
    }))
    .query(async ({ ctx, input }) => {
      const sessions = await ctx.db.session.findMany({
        where: { userId: ctx.session.user.id },
        take: input.limit + 1,
        cursor: input.cursor ? { id: input.cursor } : undefined,
        orderBy: { createdAt: 'desc' },
      });
      const hasMore = sessions.length > input.limit;
      return {
        items: sessions.slice(0, input.limit),
        nextCursor: hasMore ? sessions[input.limit - 1].id : undefined,
      };
    }),

  // Mutations — write data
  create: protectedProcedure
    .input(createSessionSchema)
    .mutation(async ({ ctx, input }) => {
      return ctx.db.session.create({
        data: { ...input, userId: ctx.session.user.id },
      });
    }),

  end: protectedProcedure
    .input(z.object({ id: z.string(), summaryNotes: z.string() }))
    .mutation(async ({ ctx, input }) => {
      // Verify ownership before mutation
      const session = await ctx.db.session.findUnique({ where: { id: input.id } });
      if (!session || session.userId !== ctx.session.user.id) {
        throw new TRPCError({ code: 'FORBIDDEN' });
      }
      return ctx.db.session.update({
        where: { id: input.id },
        data: { endedAt: new Date(), summaryNotes: input.summaryNotes },
      });
    }),
});

Database Patterns (Prisma)

Query Best Practices

// ✅ Select only needed fields (performance)
const user = await db.user.findUnique({
  where: { id },
  select: { id: true, email: true, name: true },  // NOT select: undefined (fetches all)
});

// ✅ Avoid N+1 — use include for relations
const sessions = await db.session.findMany({
  where: { userId },
  include: {
    notes: { orderBy: { createdAt: 'desc' }, take: 1 },
    _count: { select: { crisisEvents: true } },
  },
});

// ❌ N+1 — never do this
const sessions = await db.session.findMany({ where: { userId } });
for (const session of sessions) {
  session.notes = await db.sessionNote.findMany({ where: { sessionId: session.id } }); // N queries!
}

// ✅ Cursor-based pagination for large tables
const items = await db..({
  : limit + ,
  : cursor ? { : cursor } : ,
  : { :  },
});

Security Practices

Input Validation — Always Use Zod

// Every API input goes through a Zod schema
const schema = z.object({
  email: z.string().email().max(254),
  age: z.number().int().min(18).max(120),
  content: z.string().min(1).max(10000).trim(),
});

// tRPC validates automatically — but also validate at data boundaries
const parsed = schema.safeParse(rawInput);
if (!parsed.success) throw new TRPCError({ code: 'BAD_REQUEST', cause: parsed.error });

Authorization — Always Check Ownership

// Every resource mutation must verify the user owns it
async function ensureOwnership(resourceId: string, userId: string) {
  const resource = await db.session.findUnique({
    where: { id: resourceId },
    select: { userId: true },
  });
  if (!resource || resource.userId !== userId) {
    throw new TRPCError({ code: 'FORBIDDEN', message: 'Access denied' });
  }
}

Rate Limiting

// lib/rate-limit.ts
import { Ratelimit } from '@upstash/ratelimit';
import { redis } from './redis';

export const sessionCreationLimit = new Ratelimit({
  redis,
  limiter: Ratelimit.slidingWindow(5, '1h'),  // 5 sessions per hour
  prefix: 'rl:session:create',
});

// In your router:
const { success } = await sessionCreationLimit.limit(ctx.session.user.id);
if (!success) throw new TRPCError({ code: 'TOO_MANY_REQUESTS' });

Error Handling

// Custom error classes
import { TRPCError } from '@trpc/server';

// Use appropriate error codes
throw new TRPCError({ code: 'NOT_FOUND', message: 'Session not found' });
throw new TRPCError({ code: 'FORBIDDEN', message: 'Access denied' });
throw new TRPCError({ code: 'BAD_REQUEST', message: 'Invalid input' });
throw new TRPCError({ code: 'INTERNAL_SERVER_ERROR', message: 'Something went wrong' });

// Always log server errors with context (not BAD_REQUEST — those are expected)
logger.error('Session creation failed', {
  userId: ctx.session.user.id,
  error: error.message,
  stack: error.stack,
});

Caching Strategy (Redis)

// lib/redis.ts
import { Redis } from '@upstash/redis';
export const redis = Redis.fromEnv();

// Pattern: cache with TTL, invalidate on mutation
async function getUserProfile(userId: string) {
  const cacheKey = `profile:${userId}`;
  const cached = await redis.get<UserProfile>(cacheKey);
  if (cached) return cached;

  const profile = await db.psychProfile.findUnique({ where: { userId } });
  await redis.setex(cacheKey, 300, profile);  // 5-minute TTL
  return profile;
}

// Invalidate on update
async function updateProfile(userId: string, data: Partial<UserProfile>) {
  await db.psychProfile.update({ where: { userId }, data });
  await redis.del();  
}

Common Backend Commands

npx prisma migrate dev          # Create and apply migration
npx prisma migrate deploy       # Apply migrations in production
npx prisma studio               # Visual DB explorer
npx prisma generate             # Regenerate Prisma client after schema change
npx prisma db seed              # Run seed file
npm run type-check              # Verify types server-side