Md Anish Ahamad

Backend Engineer · AI Evaluation and Agent Reliability

Greater Noida, India · +91 70339 35778 · [email protected]
github.com/anishraja04 · linkedin.com/in/md-anish-ahamad · anishahamad.com

Summary

Backend engineer who builds agent systems and the harnesses that catch them being wrong. Built a two-agent debugging system that resolved 11 of 11 incidents where a single-prompt baseline resolved 0, by refusing any fix until the full test suite passes. Run two live products used by exam candidates across India. Currently paid to author the tasks that evaluate AI coding agents.

Skills

Languages: Python, Java, SQL, JavaScript
Backend: FastAPI, Spring Boot, Django REST Framework, REST API design, Spring Data JPA, JWT authentication
Data: PostgreSQL, MySQL, MongoDB, ETL
Infrastructure: Docker, Docker Compose, GitHub Actions, Terraform, AWS EC2, AWS VPC, Nginx, Linux, Git
AI and agents: AI evaluation, benchmark authoring, agent architecture, LangGraph, MCP, RAG, FAISS, prompt engineering, structured output

Work Experience

AI Coding Task Author

Jul 2026 – Present

Handshake · Remote

Author realistic software engineering tasks used to evaluate AI coding agents, writing the issue description, a reference implementation and the automated tests that decide whether a fix is correct.
Grade AI-generated code for correctness and adherence to engineering practice, distinguishing output that is right from output that only reads as right.
Build and validate tasks against real repositories using Git and GitHub workflows.

CFD Engineer – AI Task Designer

Jul 2026 – Sep 2026

Alignerr · Remote

Designed AI evaluation tasks in computational fluid dynamics with precise problem statements, reference solutions and deterministic checking, using OpenFOAM.
Assessed AI-generated engineering answers for technical accuracy against defined criteria.

LLM Post-Training Intern

Jan 2026 – Apr 2026

Ethara AI · Gurugram, India

Refined datasets and engineered prompts to improve large language model output quality.
Analysed failure cases and validated model behaviour to assess deployment readiness.

Products and Projects

exactsizetool.com — document tooling for exam forms

exactsizetool.com · live

Built and run a suite of roughly 90 free browser-based tools that resize photos, signatures and documents to the exact pixel and kilobyte limits Indian government exam portals require.
All processing runs client side, so candidate photos and signatures never leave the device and no server holds personal documents.
Added an assistant that reads the raw instruction text from an exam portal and configures the correct tool and values automatically.

gateguide.co.in — GATE preparation books

gateguide.co.in · live

Built and run a study business: six books covering CSE, ECE, Civil, Data Science and AI, Electrical and Mechanical, totalling 5,732 questions with worked solutions and 60 full mock tests.
Derived topic weighting by analysing real past papers question by question rather than following syllabus order; published 80 free guides as the acquisition channel.

Incident Engineer — agentic root-cause debugging

github.com/anishraja04/incident-engineer

Built a two-agent system that diagnoses and fixes software incidents: a triage agent forms hypotheses from logs, an investigator agent uses tools (read, grep, run-tests, sandboxed Python) to locate the cause and apply a fix.
Designed a verification gate so a fix is only accepted once the full test suite passes, preventing the confident but incorrect patches single-prompt approaches produce.
Measured against a deterministic harness over 11 incidents: 11 of 11 resolved versus 0 of 11 for the baseline, at roughly 5 minutes and 0.0079 USD per task.

KaamSathi — team task manager

github.com/anishraja04/team-task-manager

Full-stack task manager on Django REST Framework and PostgreSQL with JWT authentication and a React single-page front end.
Implemented role-based access control at two levels, platform wide and per project, so authorisation checks both permission and project membership.
Deployed on Railway with the API, front end and database as separate services.

Cloud-Native DevOps deployment platform

github.com/anishraja04/cloud-native-devops

Provisioned AWS infrastructure as code with Terraform: VPC, public subnet, security groups and an EC2 instance.
Containerised the application behind an Nginx reverse proxy, with GitHub Actions building the image and deploying on every push.

Education

Bachelor of Technology (B.Tech), Computer Science and Engineering

Jul 2022 – Jun 2026

Noida Institute of Engineering and Technology · Greater Noida, India

CGPA 8.42 out of 10, with Honors.

Certifications

Data Structures — University of California San Diego Mar 2024
Introduction to Artificial Intelligence — IBM Jan 2023
Frontier Engineering Challenge 2026, Participant — micro1 2026

A note on the PDF. The downloadable version is built to parse cleanly in applicant tracking systems: one column, no tables, no text boxes, nothing in the page header, real Unicode bullets and embedded fonts. It is checked after generation by extracting its own text layer and comparing it word by word against the source — if that check fails, the file is not written.