Software Engineer · Backend, Full-Stack & AI

Ashish Yadav

Engineering high-throughput cloud backends, scalable full-stack platforms, and resilient production AI systems.

I build end-to-end software designed for real-world traffic. Experienced across Go, Python, TypeScript (Next.js/React), Flutter, and AWS — from architecting serverless backends serving 100K+ active users at <200ms p95 latency to shipping 18 mobile apps with 250K+ downloads (4.9★) and autonomous AI agent infrastructure.

Scale / users

100K+

Active Daily Users

Mobile / apps

250K+

4.9★ (18 Apps Shipped)

Latency / target

<200ms

p95 Production Target

View Resume Explore Projects My Creative Side/atelier

mini-drop/engine/stream_dedup.go

Load:

1	package engine
2
3	// bufferPool reuses 32KB byte slices to eliminate GC pressure
4	var bufferPool = sync.Pool{
5	New: func() any { buf := make([]byte, 32*1024); return &buf },
6	}
7
8	// WriteChunk hashes streaming reader chunks without heap allocation
9	func (e *Engine) WriteChunk(r io.Reader) (string, error) {
10	buf := bufferPool.Get().(*[]byte)
11	defer bufferPool.Put(buf)
12	h := sha256.New()
13	if _, err := io.CopyBuffer(h, r, *buf); err != nil {
14	return "", err
15	}
16	return hex.EncodeToString(h.Sum(nil)), nil
17	}

High-Throughput Backend & Cloud Infrastructure

Zero-allocation streaming, transactional database state machines, and sub-200ms real-time APIs built for production traffic.

Production Scale100K+ Active Users

App Downloads250K+ (4.9★ Rating)

API Latency<200ms p95 SLA

Production Scale: 100K+ Active Users·App Downloads: 250K+ (4.9★ Rating)

API Latency: <200ms p95 SLA·Mobile Ecosystem: 18 Production Apps·Competitive Coding: 430+ Solved

Work Experience

Building and shipping production backends, full-stack apps, and AI products for real users.

Active Production Role · Remote

Resume (PDF)

ASOasis Tech

Remote

Software Developer Intern (Backend, Full-Stack & AI)

Aug 2025 — Present

Owned end-to-end product engineering across cloud backends, mobile apps, and AI pipelines serving 100K+ active users and 250K+ total downloads.

5 Key Deliverables100K+ Scale

Deliverable 01 · Go Microservices & Recommendation Engine100K+ Users · <200ms SLA

100K+ Active Users — Scalable Backend & Recommendation Engine

What I Built & Shipped:

Built a serverless Go backend serving 100K+ active users with <200ms real-time AI chat, a Redis/MySQL recommendation engine (Gorse), and consistent hashing (FNV-32a) for model routing.

Execution StatusLive in Production

Target Impact100K+ Users · <200ms SLA

Technologies Used:

GoAWS LambdaRedisMySQLGorseFNV-32a Hashing

Technical Skills & Stack

All-in-one market stack verified across production backends, cloud infrastructure, full-stack web, and AI systems.

5 Core Engineering Domains

Languages

Core Fundamentals

Go (Golang)PythonTypeScript / JavaScriptDartSQL

Backend & Cloud

Scalable Infrastructure

AWS (Lambda, Step Functions, EventBridge, DynamoDB, S3)FastAPINode.jsREST APIsWebSockets & SSEDockerGitHub Actions CI/CD

Frontend & Mobile

Modern Interfaces

Next.js 16React 19Tailwind CSSFlutter (iOS & Android)React Flow

Databases & Caching

Data Persistence

PostgreSQLRedisMySQLDynamoDBSQLiteBigQuery

AI & Machine Learning

Production Intelligence

LLM OrchestrationAI AgentsRAG & Vector SearchGoogle Gemini / OpenAI APIsLoRA Fine-Tuning (MLX)Faster-Whisper STT

Synchronizing Execution Pipeline...

Featured Engineering Projects

Full-stack platforms, AI agent systems, and distributed backends built from scratch.

AgentOS — Autonomous AI Agent Platform

Flagship

Full-Stack & Cloud AI · Go + Next.js + AWS

GovernanceHuman-in-the-Loop Safe

Code

A full-stack platform to run, schedule, and monitor autonomous AI agents safely in the cloud with human approval gates for sensitive actions.

Key Features & Architecture

01Serverless Workflow Engine

Uses AWS Step Functions (waitForTaskToken) to pause AI agents whenever they request sensitive permissions and resume instantly after human approval.

02Reliable Database Transactions

Implements the Transactional Outbox pattern in PostgreSQL (pgx/v5) with optimistic locking so no task events are ever lost.

03Instant Cache Updates

Caches agent permissions and long-term user memory in Redis with instant write-through invalidation.

04Live Mission Control UI

Next.js 16 and React 19 dashboard with real-time step receipts, cron scheduling via EventBridge, and one-click approvals.

GoNext.js 16AWS Step FunctionsPostgreSQLRedisAWS CDK

Release Flight Control — Staged Deployment & Rollback System

Flagship

Backend & DevOps Control Plane · Go + Next.js

AutomationReal-Time Auto-Rollback

Code

A deployment management platform that rolls out app updates in safe stages (1% → 5% → 25% → 100%), monitors live error metrics, and triggers automatic rollbacks when bugs happen.

Key Features & Architecture

01Automated Metric Guardrails

Background Go workers continuously monitor live app health metrics and flag releases if success rates drop below threshold.

02Dynamic Feature Flag Control

Integrates with Firebase Remote Config via clean adapter interfaces to update rollout percentages or revert versions instantly.

03Live Dashboard Streaming

Streams real-time deployment status and immutable audit logs to the Next.js frontend using Server-Sent Events (SSE).

04Safe Concurrency

Built with Go goroutines, channels, sync.RWMutex, and graceful OS signal shutdown (SIGINT/SIGTERM).

GoNext.jsPostgreSQLServer-Sent Events (SSE)FirebaseTypeScript

SketchTalk — Voice-to-Architecture Whiteboard

Full-Stack AI Product · FastAPI + Gemini + React

Latency<1ms Fast-Path Edits

Code

An interactive AI whiteboard where engineers speak system architectures out loud and watch validated diagrams build and update live on canvas.

Key Features & Architecture

01Hybrid Voice Pipeline

Combines Web Speech captions, faster-whisper audio streaming, and RapidFuzz phonetic correction to fix spoken tech jargon ("post gress" → PostgreSQL).

02Zero-Latency Rule Engine

Executes common voice commands ("add Redis", "replace RabbitMQ with Kafka") in <1ms via deterministic rules, calling Google Gemini only for complex sentences.

03Hallucination Guardrails

Strict Pydantic semantic validator blocks broken connections, self-loops, and invalid graph edits (0% unsafe mutations).

04Interactive Studio UI

Built with Next.js 16 and React Flow v12 featuring smart node placement, O(1) undo/redo, and Mermaid/JSON export.

Python (FastAPI)Google GeminiNext.js 16React FlowFaster-WhisperWebSockets

Distributed Fault-Injection Testing Engine & Mini-Dropbox

Core Backend & Reliability · Python + Go

Throughput2,000 req/s (4x Gain)

Code

High-performance backend testing and storage systems: a deterministic fault-injection engine for microservices and a concurrent file storage server in Go.

Key Features & Architecture

01Distributed Bug Detection

Simulates network drops, delays, and race conditions across 4 microservices, catching 15/15 concurrency bugs with 0 false alarms across 5,000 workflows (178k events/sec).

02Root-Cause Trace Slicer

Automatically shrinks 40-event crash logs down to the exact 5 causal events that caused the failure, with 100% reproducible SHA-256 replay.

03High-Speed File Storage (Mini-Dropbox)

Built a content-addressed storage engine in Go using sync.Pool 32KB buffers, boosting concurrent throughput 4x to 2,000 req/s and cutting RAM usage by 43%.

04AI Eval & Fine-Tuning Suite (AgentGym)

Created 3 sandboxed AI testing environments backed by 106 passing unit tests and ran local LoRA fine-tuning on Apple Silicon (mlx-lm).

GoPython 3.12SQLiteHypothesis FuzzingPytestsync.Pool

Let’s Build Great Software Together

Available immediately for internships and full-time Software Engineering roles (Remote or On-site).

Active Across Global Timezones

Friends & Fellow Builders

The Tech Fam & Friends I Build With

Hover on any profile to meet the brilliant engineers, analysts, and creators in my corner.

11 Amazing Friends

Rendering Vector Projections...

Available for SDE Roles·Immediate Joiner

Let’s build something impactful together.

Open for full-time Software Engineer positions across backend, full-stack, and AI systems. Have an opportunity or want to collaborate? Reach out directly.

Collaboration TimelineFast-Track Onboarding

01<2h SLA

First Contact

Drop an email or call. Fast async intro and role alignment.

02Deep Dive

Technical Sync

System design discussion, stack fit, and live code walkthrough.

03Day 1 Ready

Ship to Prod

Zero lead time. Immediate start building production systems.

ashish.devv2026@gmail.com

Send Email

TEL:+91 875-008-4388

Call Directly

Direct Profiles & Verification

Explore coding history, production apps, and verified problem-solving track record.

LinkedInashishyadav001

GitHub@Ashish5180 · Production Repos

LeetCode ProfileDeveloper-AshishYadav · Algorithms & DSA

Codolio Profile430+ Problems Solved · Competitive Streak

Resume PDFDownload Updated Curriculum Vitae

Designed & Engineered by Ashish Yadav · Software Engineer (Backend · Full-Stack · AI)

Connecting Telemetry GridEstablishing low-latency WebSocket connection to global edge nodes...

Ashish Yadav

High-Throughput Backend & Cloud Infrastructure

Ashish Yadav

High-Throughput Backend & Cloud Infrastructure