Portfolio · Bengaluru

Mohammed Aman Khan

AI-first backend, platform & product-minded execution

Senior Software Engineer

I am a self-made engineer who learned by building, shipping, breaking, and improving real systems. My work sits at the intersection of backend depth, platform thinking, and applied AI - from payment infrastructure and cloud architecture to LLM features that need to be fast, grounded, observable, and safe in production.

I like going deep, learning fast, and owning hard problems end to end - especially where AI meets real product and infrastructure constraints.

Get the highlights.

AI-first, not AI-only

AI-first backend, platform & product-minded execution

See experience

Built for real constraints

Security, tenancy, reliability, and cost awareness at scale

View projects

LLMs that survive production

Production LLM systems with retrieval, evals, and tool calling

Explore skills

About me

How I work

I am based in Bengaluru and have grown by being relentlessly hands-on. I am strongest in backend and cloud systems, but I am most useful when a problem cuts across APIs, infrastructure, product behavior, and AI workflows at the same time.

I stay close to new models, tooling, and developer workflows, but I am not interested in AI demos that fall apart in production. I care about evals, latency, tenancy boundaries, grounding, rollout safety, and whether the feature actually helps the user.

Strong delivery starts with clarity on business constraints, customer impact, and operational burden. I ask direct questions, document tradeoffs, and bias toward systems that teams can understand, operate, and extend without heroics.

Core stack

Technical skills

Languages, platforms, and AI capabilities from years of shipping production backends and client applications. The snapshot matches what you will see expanded in the categories below.

Stack snapshot

  • Python
  • TypeScript
  • JavaScript
  • React
  • Next.js
  • Node.js
  • FastAPI
  • Express
  • AWS
  • Docker
  • Kubernetes
  • PostgreSQL
  • GraphQL
  • DynamoDB
  • Firebase
  • Kafka
  • OpenAI
  • Google Cloud
  • Flutter
  • Git

Languages

  • Python
  • TypeScript
  • JavaScript
  • HTML
  • CSS
  • Dart (beginner)
  • Bash (beginner)

Frameworks

  • FastAPI
  • Next.js
  • Express
  • Serverless
  • Node.js
  • React
  • Redux Toolkit
  • Flutter
  • GetX

Full-stack

  • REST & OpenAPI APIs
  • GraphQL (AppSync)
  • React + TypeScript UIs
  • Serverless & event-driven backends
  • Auth flows (Cognito, SSO)
  • End-to-end feature ownership

Tools & data

  • Firestore
  • PostgreSQL
  • Vespa DB
  • Kafka
  • Docker
  • Kubernetes (beginner)
  • Git

AWS

  • EKS
  • Lambda
  • API Gateway
  • AppSync
  • Cognito
  • DynamoDB
  • S3
  • CloudWatch
  • IAM (beginner)

Platforms

  • Firebase
  • GCP (beginner)
  • GitLab CI/CD (beginner)
  • Jenkins (beginner)

AI & ML

  • Azure OpenAI
  • OpenAI-compatible APIs
  • Google Gemini
  • Prompt engineering & evals
  • Embeddings & vector search
  • Retrieval-Augmented Generation (RAG)
  • Agentic orchestration & tool calling
  • LLM + OpenAPI backends
  • Keras (classical ML)

Leadership & craft

  • Technical leadership
  • Ownership & delivery
  • Documentation
  • System design
  • Cross-functional collaboration

Work history

Experience

Senior-level ownership across payment platforms, AWS serverless and Kubernetes stacks, and applied AI - architecture, security, delivery automation, and measurable outcomes.

Juspay

Software Development Engineer II

November 2024Present · Bengaluru

  • Merchant cost observability (Hyperswitch / IC++)

    • Designed and shipped self-serve analytics for interchange-plus-plus pricing: consolidated fee breakdowns, drift detection, and tuning levers so merchants and internal finance could reason about margin without ad hoc extracts.
    • Partnered with product on metric definitions and data contracts so dashboards stayed trustworthy as pricing rules evolved.
  • Telemetry-based anomaly detection and cost optimization

    • Owned the end-to-end design of a scoring service over live card payment streams: prioritization, root-cause hints, and operator-facing remediation steps.
    • Contributed to roughly a ten percent reduction in card processing cost quarter over quarter by steering configuration toward healthier rails and fewer misrouted transactions.
  • Microservices and Kubernetes on AWS EKS

    • Extended internal payment microservices for sustained high concurrency: clearer service boundaries, safer rollouts, and failure isolation on critical paths.
    • Aligned capacity planning, metrics, and tracing with availability goals for merchant-facing APIs and downstream processors.
  • Internal agentic assistant (RAG + tool calling)

    • Built multi-step orchestration combining Vespa-backed retrieval-augmented generation, OpenAPI-bound tools, and per-user, per-tenant authorization.
    • Grounded responses in internal APIs and documentation to shorten incident triage and reduce reliance on static runbooks.

Softway

Software Engineer II

August 2022November 2024 · Bengaluru

  • Culture+ Counter: enterprise survey and analytics platform

    • Led end-to-end engineering for surveys and analytics on the Culture+ model: metric design, data model, and reporting that held up in leadership and audit conversations.
    • Delivered workflows that connected measurement to follow-up actions for HR and executives, not only static dashboards.
  • Event-driven survey pipeline (DynamoDB, Lambda)

    • Cut median submission latency about thirty percent by shifting heavy aggregation to DynamoDB-triggered workers with asynchronous, horizontally scaled processing.
    • Separated hot write paths from batch backfills so user-facing latency stayed stable under load.
  • GraphQL layer and deployment engineering

    • Introduced AWS AppSync where typed queries and subscriptions fit the product, integrated with the existing Lambda and DynamoDB data plane.
    • Reduced manual production steps about seventy percent via scripted pipelines, environment promotion rules, and repeatable release checks.
  • Identity, authorization, and data isolation (Cognito)

    • Hardened authentication with Amazon Cognito and enforced user-scoped access patterns across sensitive DynamoDB datasets.
    • Constrained privileged operations through reviewed code paths to lower the blast radius of configuration mistakes.
  • ML-backed insight ranking

    • Implemented a ranking pipeline using pairwise comparison and K-means clustering so generated insights aligned with product priorities.
    • Packaged ranked output for sales and customer success as a repeatable narrative in enterprise deals.
  • React platform: API clients and cross-cutting concerns

    • Established Axios interceptor patterns for auth headers, structured errors, and retry policy across a growing React and TypeScript codebase.
    • Gave feature teams consistent contracts as REST and GraphQL surfaces expanded.
  • Culture+ Admin: greenfield B2B CRM

    • Owned architecture and delivery for a new CRM covering onboarding, entitlements, and product allocation from first commit through production traffic.
    • Supported roughly twenty percent revenue growth by replacing multi-day manual onboarding with guided, auditable automation.
  • Enterprise SSO and login performance

    • Integrated SSO and third-party identity providers for large tenants with clear session and token lifecycle handling.
    • Improved median login time about fifty-five percent by streamlining the primary authentication path and removing redundant steps.

Baliza

Software Engineer

August 2021May 2022 · Remote

  • Velor: Flutter client for live video and calling

    • Built cross-platform streaming and in-app video calling on ConnectyCube with emphasis on resilience on residential WiFi and variable devices.
    • Profiled and tuned bitrate, buffering, and reconnection to reduce frozen frames during sustained sessions.
  • Realtime control plane (Firebase)

    • Specified compact control messages for recording, lullabies, and alerts to keep payloads small and latency predictable.
    • Used Firebase Realtime Database for ordered, sub-second signaling and remote device commands.
  • Street Runner: Web3 marketing stack (React)

    • Delivered a multilingual React front end with MetaMask connect, event listings, and regional campaign scheduling for a crypto racing title.
    • Structured CMS-friendly content so marketing could ship locales without per-market engineering tickets.

Selected work

Projects

Production-adjacent work and public repos — NLP and voice, serverless SaaS, mobile, and full-stack experiments.

Samantha

  • Python
  • Keras
  • NLP
  • pyttsx3
  • Speech recognition

Voice-first assistant: on-the-fly phrase learning, task execution from speech, and scripted automation with device-level control—research-oriented NLP side project.

  • Keras DNN + NLP for core conversational routing and intent handling.
  • Low-latency voice loop: speech recognition, pyttsx3, and multi-threaded I/O.
  • Incremental phrase learning plus keyboard/mouse hooks for repeatable desktop workflows.

Culture+ Counter

  • Python
  • AWS Lambda
  • DynamoDB
  • Cognito
  • TypeScript
  • React
  • Redux

Enterprise survey and analytics product on the Culture+ framework—measurable culture signals and action-ready reporting for HR and leadership.

  • Serverless data plane: Cognito-scoped access, DynamoDB-backed survey lifecycle.
  • React + Redux admin and author tooling for launch, monitoring, and operations.

Culture+ Admin

  • Node.js
  • TypeScript
  • AWS Lambda
  • DynamoDB
  • Cognito

Greenfield CRM for the Culture+ suite: client onboarding, product allocation, and day-two operations at scale.

  • Built from scratch for onboarding, entitlements, and product assignment under load.
  • Tenant-scoped serverless backend: Cognito + DynamoDB isolation patterns.

E-COM

  • React
  • Express
  • JavaScript
  • Firebase Hosting

Full-stack e-commerce style site: Express API (hosted on Render) and React UI on Firebase Hosting—split deployment with a clear client–server boundary.

  • React front end with Yarn-based dev workflow; Express backend exposed for catalog and checkout flows.
  • Production-style hosting: API on Render, static and SPA assets on Firebase.

Nearby Hospital Locator

  • Flutter
  • Dart

Flutter app that reads the device location and surfaces up to thirty nearby hospitals—map-centric UX with permissions and markers.

  • Location-driven discovery with map integration for at-a-glance hospital proximity.
  • Documented flows in-repo with screenshots for onboarding and marker interactions.

Realtime team chat (Flutter)

  • Flutter
  • Dart
  • Firebase Auth
  • Cloud Firestore
  • Firebase Storage
  • FCM
  • Agora RTC

Cross-platform Flutter client for Slack- and WhatsApp-style messaging: Firebase-backed identity and data, Agora for voice and video, and modern chat UX in one app.

  • Firebase Auth for accounts and session lifecycle; Firestore for conversations, membership, and ordered message streams; Storage for media; FCM for push.
  • Agora RTC for in-app voice and video calls tied to threads and participants—no separate dialer flow.
  • Product surface inspired by Slack and WhatsApp: direct and group threads, channels or spaces as needed, presence, typing and read receipts, reactions, and rich media previews.

Chatbot

  • Python
  • Keras
  • Flask

Small Keras-backed conversational API: trained weights bundled as HDF5, served behind a lightweight Python app for inference.

  • Neural model packaged as chatbot.h5 with Flask entrypoints for request/response handling.
  • JSON-driven training data and Heroku-style deployment artifacts (Procfile, runtime).

Foundation

Education

Birla Institute of Technology and Science, Pilani

Master's in Artificial Intelligence & Machine Learning

Bengaluru, India

April 2026Present

CMR Institute of Technology, Bengaluru

Bachelor of Engineering in Computer Science

Bengaluru, India

July 2018July 2022

Contact

Hiring? Let's talk.

Open to full-time backend, platform, and AI-adjacent engineering roles. Email or LinkedIn for the quickest reply - or book a time. Resume below.

© 2026 Mohammed Aman Khan · Bengaluru, India