Abdul Salam Issahaku

Software Developer

I build open digital infrastructure at the intersection of sustainability, AI, and data interoperability. My work spans geospatial platforms, AI-powered decision tools, and open sustainability analytics across the Global South.

Projects

Green-Irrigate Ghana

Built a map-centric platform that helps agricultural extension officers and farmers assess solar irrigation feasibility for smallholder farms.

Next.js · FastAPI · Python · PostGIS

Agrospectrum Website.

A cooperate website for Agrospectrum Ltd. using Next.js, Tailwind, Figma, and Strapi.

improving brand presence and client engagement

Next.js · Tailwind CSS · Strapi

Fiber Backbone Network Manager

Built at Broadspectrum.

A geospatial management system for visualizing a subsidiary's fiber infrastructure, mapping points of presence and routes using Next.js and React-Leaflet

Next.js · React-Leaflet · MongoDB

Mindgraph Research

An AI that transforms research papers into knowledge graphs, helping users quickly understand complex fields, identify seminal works, and uncover novel research opportunities.

FastAPI · Langchain · Gemini · Next.js · Python

Tarvan.AI

An NFT gaming marketplace. Integrated FastAPI-based AI agent for real-time price updates driven by market activity.

Next.js · FastAPI · Python

Learning Management System

An LMS supporting online course delivery, discussion forums, and video conferencing for 100+ users. Executed complex state management using React Context API.

React.js · Node.js · MongoDB · Firebase

Inventory Manager

A full-stack inventory management system, providing real-time tracking of inventory, sales, and user accounts through a responsive interface.

Next.js · TypeScript · Prisma

Latest Blogs

All blogs →

Today I Learned

Aug 7, 2026

Mean Reciprocal Rank (MRR)

Top-1 accuracy is all-or-nothing: a ranker that puts the right answer second scores exactly zero. MRR grades the whole ranking instead; take the reciprocal of the position where the correct answer landed (1st → 1.0, 2nd → 0.5, 4th → 0.25, missing → 0) and average it across all queries. Handy intuition: 1 ÷ MRR ≈ where the right answer typically sits in the list. I met it while comparing a lexical name-matcher against a sentence-transformer for ranking schema-field suggestions: their top-1 counts were close, but MRR (0.48 vs 0.60) revealed the model was missing higher up the list, which is the metric that actually matters when a human confirms answers from a ranked list

Aug 6, 2026

Thresholded abstention

A classifier should decline to predict when its confidence falls below a threshold, trading coverage for accuracy. The modern literature calls it selective prediction / selective classification (Geifman & El-Yaniv, 2017), and related framings include open-set recognition and out-of-distribution detection, all variants of "know when none of your known answers apply.

All TILs →

About

I'm a software developer and open source contributor building infrastructure that makes environmental data trustworthy and actionable.

My work sits at the intersection of sustainability technology, AI systems, and data interoperability. I build full-stack systems across agritech, fintech, and AI-driven platforms. I work primarily in React, Next.js, TypeScript, FastAPI, and Python, with PostGIS for spatial data and LangChain for AI integration.

I care about building systems that hold up under real conditions, not just technically, but epistemically.

Skills

Programming Languages
Python, Javascript
Frontend
React.js, Next.js, TypeScript, Tailwind CSS, Responsive Design.
Backend
Node.js, FastAPI, tRPC, Prisma
Databases
MySQL, PostgreSQL, MongoDB, SQLite
Cloud & DevOps
AWS, Docker, Git, DigitalOcean, Render
Web3
Wagmi, Smart Contract Integration
Other
Technical Writing

Experience

Open Source Contributor

Open Sustainable Technology Germany • October 2025 - Present

Contributing to a directory and analysis of the open source ecosystem covering climate change, sustainable energy, biodiversity, and natural resources. Refactored a monolithic codebase into modular components and contributed to the project's analytics dashboard.

Software Developer

Broadspectrum Limited Ghana • November 2023 - Present

I work across internal and external agritech systems, owning the frontend development of an agent onboarding portal and agribusiness admin platform, and building a fiber backbone network management system with geospatial visualization. I designed a schema-driven dynamic data collection engine that lets non-technical users create complex forms with conditional logic, and spearheaded the integration of generative AI into the company's products. I also built a fully responsive subsidiary website using Next.js, Tailwind, and Strapi. Owned the frontend development of a multi-tenant agricultural subsidy management platform. Which handles farmer registration, input distribution, program management, marketplace transactions, financial reporting, and field monitoring across multiple organizations and languages.

Frontend Developer

SoftinHub • December 2023 - May 2025

Contributed to the incremental frontend development of an online retail marketplace, delivering feature updates and UI improvements that enhanced the shopping and checkout experience.

Technology Instructor Intern

OpenLabs Ghana • August 2022 - January 2023

Taught an introductory software development course using Python, guiding students through fundamentals with a 100% pass rate.