WIT Summer Mentorship Programme 2026
The Centre for Water Informatics and Technology (WIT) invites undergraduate students to apply for its Summer Mentorship Program (SMP) 2026. This one-month program offers students the chance to work on innovative projects that contribute to WIT's research agenda. Participants will gain hands-on experience, mentorship from WIT faculty and staff, and a stipend upon successful completion. The program this year will run from July 01, 2026, to July 31, 2026. Apply by June 10, 2026, to be considered for the program.
Duration: The WIT SMP 2026 will take place from July 01, 2026 to July 31, 2026. Applications will be accepted till June 10, 2026.
Eligibility: Undergraduate students from all universities are eligible to apply. Depending on the nature of the work, students of only some specific disciplines may be asked to apply for a particular project. Details for each individual project are published prior to the call for applications. Note that while students from outside Lahore are also eligible to apply, they must arrange their own accommodation for the project duration.
Selection Procedure: Students apply for specific projects in the shared project list. Student applications will be reviewed by the WIT faculty, and successful applications will be notified accordingly. Each project is assigned to a mentor from the WIT staff who will supervise the student during the program.
Compensation: Each student will receive a stipend of PKR 20,000, which will be awarded after successfully completing the program.
Project Presentation and Report: At the end of the program, the student will present their work in a joint session and submit a written report. These will be evaluated by the WIT faculty to rank the projects for the best project award.
Best Project Award: The winner of the best project award will receive support to continue the project for an additional period of one month. In addition to the stipend for the extended duration, the student-mentor pair will also have access to a small grant of PKR 50,000 that can be utilized for travel, training, and purchase of materials and equipment related to the project activities.
Past SMP Projects: The SMP takes place every year. Program and project details for previous years can be found here.
How to apply: Interested candidates must send their applications by email titled "WIT SMP 2026 Application" to hrwit@lums.edu.pk. The email must include the applicant's CV, academic transcripts, and a cover letter (no more than 300 words) detailing why the applicant is best suited for the proposed work. In the email, the applicant must clearly specify the project for which the application is being submitted. A maximum of two projects can be applied for (with preferences 1 and 2 clearly mentioned).
Below are details for the available projects in summer 2026.
Projects for WIT SMP 2026
Available projects: A total of 8 projects are being offered. The list of projects for the year 2026 can be found below:
- Exploring watershed models for simulating streamflow dynamics
- Benchmarking and Intelligent Calibration of Soil Moisture Sensors Using Tensiometer and Gypsum Block Systems
- Edge-AI for Disaster Resilience: Deploying Local IoT Infrastructure and Open-Source LLMs for Autonomous Flood Early Warning Triggers
- Mapping Pakistan’s Agricultural Trade: Building a Provincial Statistics and Global Commodities Database for FABLE
- Spatio-Temporal Groundwater Mapping of the High Plains Aquifer
- Bias Correction and Local Calibration of Remote Sensing–Based Gridded Precipitation and Runoff Datasets for Hydrological Applications
- Data-Driven Streamflow Prediction for the Ravi River at Jassar using Statistical and Machine Learning techniques
- Community, Control and Conservation: A Comparative Study of Indigenous Water-Sharing Systems in Chitral and Colonial Irrigation Institutions in Punjab
Exploring watershed models for simulating streamflow dynamics
Mentor: Nihal Ahmed, Research Assistant, WIT.
Majors: All science and engineering majors with the requisite knowledge can apply.
Necessary Skills: Fundamentals of GIS, basic Python knowledge, and the ability to conduct independent technical research.
Description: Physical hydrological models such as SWAT and HEC-HMS serve a wide number of applications including flood early warning systems (FEWS), water quality management, and predicting the environmental effects of land use and agricultural practices. Traditionally FEWS rely on static thresholds. If the reported stream level is above a certain threshold, an alarm is triggered. However, providing the downstream communities with actionable lead time requires predictive models.
In this research project, students will build a predictive watershed model that forecasts stream water levels and discharge rates, simulating how a local watershed responds to various weather inputs. By sourcing historical geospatial, precipitation, and streamflow data from the US Geological Survey (USGS) National Water Dashboard, students will delineate watersheds, soil, and land-use, and calibrate their models against historical extreme events. The model will then be integrated with an API to generate forecasts after assimilating real-time observations from environmental monitoring sensors. The final product is a pipeline that enables us to model local watersheds and simulate the response to future climate scenarios or predict weather events, allowing actionable water resource management from environmental data.
Benchmarking and Intelligent Calibration of Soil Moisture Sensors Using Tensiometer and Gypsum Block Systems
Mentor: Ali Akbar, Business Development Manager, WIT.
Majors: Computer Science, Electrical Engineering, Agricultural Engineering, or comparable degree.
Necessary Skills: Basic engineering skills, sensor calibration, familiarity with data analysis tools, hands-on skills in assembling mechanical apparatus for laboratory testing, ability to conduct independent technical research.
Description: Accurate soil moisture data is critical for precision agriculture and efficient water management. Currently, the deployment of low-cost capacitive soil moisture sensors requires localized calibration using traditional gravimetric methods (drying soil samples, incrementally adding water, and recording weights). While functional, this methodology is incredibly time-consuming and prone to human error, which impacts the reliability of the calibrated data.
This project aims to develop a rapid, highly accurate laboratory benchmarking and calibration system utilizing tensiometers. By using a tensiometer to directly measure soil water tension (suction pressure), we can create a standardized baseline to calibrate capacitive sensors much faster and with higher precision.
The selected intern will work on:
- Prototype Development: Designing a laboratory test rig for benchmarking capacitive soil moisture sensors against a standard tensiometer.
- IoT Integration Roadmap: Outlining a technical roadmap to digitize the tensiometer. This will explore dual pathways: directly integrating electronic pressure sensors, or utilizing Edge AI/Vision AI to read traditional analog gauges via a camera module.
- Field Data Collection: Testing and gathering comparative data from existing sensor nodes deployed at the LUMS field site.
- Alternative Mediums (Secondary Objective): Exploring the feasibility of gypsum block-based soil moisture sensors as an additional cost-effective method for standardizing agricultural calibration.
Edge-AI for Disaster Resilience: Deploying Local IoT Infrastructure and Open-Source LLMs for Autonomous Flood Early Warning Triggers
Mentor: Ali Akbar, Business Development Manager, WIT.
Majors: Computer Science, Electrical Engineering, Network/Software Engineering, or Mechatronics.
Necessary Skills: Basic understanding of IoT architecture and edge computing, familiarity with lightweight Linux environments (Docker/Raspberry Pi), basic exposure to Python or JavaScript, and an interest in the open-source LLM ecosystem (e.g., Ollama, Llama.cpp).
Description: Flood early detection systems are frequently deployed in remote, topographically challenging locations where reliable internet connectivity is either non-existent or highly unstable. While conventional, rule-based threshold logic (e.g., simple "if-then" triggers based on water level) handles standard alerts well, it often lacks the contextual intelligence required to interpret complex, multi-sensor anomaly events or dynamic environmental patterns.
This project aims to bridge the gap between advanced generative AI and ruggedized, offline field infrastructure by exploring the deployment of localized, open-source Large Language Models (LLMs) at the network edge. Instead of routing telemetry data to external cloud services, a localized edge server will aggregate real-time sensor metrics and employ a highly optimized, small-footprint LLM to assess situational conditions, evaluate risks, and autonomously trigger adaptive warnings and alerts.
The selected intern will be responsible for:
- Edge Server Prototyping: Setting up and configuring a local edge IoT gateway containerized via Docker, utilizing open-source middleware platforms such as ThingsBoard Edge or Node-RED.
- Local LLM Integration: Exploring and benchmarking the lightweight open-source LLM ecosystem (such as Llama 3 or Phi-3 running locally via Ollama) on constrained hardware to assess inference latency and parsing accuracy.
- Intelligent Trigger Logic: Developing automation pipelines (flows/rule chains) where sensor telemetry data is structured into contextual prompts, allowing the local LLM to evaluate raw data patterns and systematically output structured alert actions, warning states, or natural-language emergency updates.
- Resilience Testing: Evaluating the prototype's end-to-end performance under fully offline, simulated network failure conditions to ensure complete runtime reliability at the field site.
By the end of the mentorship, the project will produce a functional edge-computing prototype demonstrating offline, AI-driven anomaly detection, alongside a documented technical framework for integrating localized AI into WIT's IoT development roadmap.
Mapping Pakistan’s Agricultural Trade: Building a Provincial Statistics and Global Commodities Database for FABLE
Mentor: Syeda Baseerat Fatima, WIT.
Majors: Computer Science, Data Science, Economics, Agricultural Sciences, Environmental Sciences, or related field.
Necessary Skills: Python (pandas, requests), Excel, ability to read and interpret statistical data tables.
Description: At WIT, we are developing FABLE (Food, Agriculture, Biodiversity, Land-use, and Energy)-Pakistan — a modelling tool to study food security, agricultural trade, and land-use under different climate and policy scenarios. A key challenge is the lack of a clean, consistent database linking Pakistan’s provincial crop production statistics with global trade data for key commodities.
The selected student will build a structured, reproducible data pipeline that compiles provincial agricultural statistics from official sources such as the Ministry of National Food Security and Research, Pakistan Bureau of Statistics (PBS), and provincial agriculture departments, and retrieve global trade data from FAOSTAT (Food and Agriculture Organization Statistical Database), UN Comtrade (United Nations International Trade Statistics Database), and WITS (World Integrated Trade Solution). The student will also develop a commodity concordance table mapping national crop categories to FABLE model categories, and produce a short data brief and presentation for the WIT team.
The student will gain practical experience working with real agricultural and trade datasets, writing Python data pipelines, and contributing to active modelling research at WIT.
Spatio-Temporal Groundwater Mapping of the High Plains Aquifer
Mentor: Hassaan Ahmed, PhD Student, WIT.
Majors: All science and engineering students with the requisite knowledge.
Necessary Skills: Python programming experience, basic GIS/QGIS knowledge, and a willingness to independently learn new techniques and tools throughout the project.
Description: Data-driven algorithms are only as effective as the quality, consistency, and frequency of the data provided to them. One such algorithm is Dynamic Mode Decomposition with Control (DMDc), which extracts dominant spatio-temporal patterns from snapshot data and produces reduced-order models for complex systems. This project focuses on collecting, processing, and organizing groundwater extraction data for the High Plains Aquifer (USA) to generate spatio-temporal datasets suitable for DMDc analysis. The intern will explore different data sources and preprocessing techniques to produce a data-processing pipeline that converts raw and scattered groundwater well records into structured spatial and temporal snapshots.
Bias Correction and Local Calibration of Remote Sensing–Based Gridded Precipitation and Runoff Datasets for Hydrological Applications
Mentor: Ahmad Haseeb Rabbani, PhD Student, WIT.
Majors: All science and engineering majors.
Necessary Skills: Data processing (Excel/MATLAB/Python), Basic Familiarity with Satellite based Remote Satellite products.
Description: Remote sensing–based gridded precipitation and runoff datasets provide an effective alternative for addressing the scarcity of ground observations in hydrological studies; however, these datasets often contain systematic biases arising from sensor limitations, regional climatic variability, and local hydrological conditions. This project aims to evaluate and perform bias correction of satellite and gridded precipitation and runoff products by comparing them with available ground observations and local datasets to improve their reliability under local conditions. The intern will work on data collection, preprocessing, visualization, and implementation of statistical or data-driven correction techniques using tools such as Excel, MATLAB, or Python. The project will provide hands-on experience in processing large environmental datasets, remote sensing applications, and hydrological analysis, while contributing towards improved datasets for streamflow prediction studies.
Data-Driven Streamflow Prediction for the Ravi River at Jassar using Statistical and Machine Learning techniques
Mentor: Ahmad Haseeb Rabbani, PhD Student, WIT.
Majors: All science and engineering majors.
Necessary Skills: Data processing (Excel/MATLAB/Python), basic understanding of machine learning concepts, data analysis, and familiarity with hydrological datasets.
Description: Streamflow prediction plays a critical role in water resources management, flood forecasting and agricultural planning. Traditional hydrological models often require extensive physical data and complex parameterization, while data-driven approaches provide an efficient alternative by learning relationships directly from historical observations and environmental variables. This project aims to develop data-driven models for streamflow prediction of the Ravi River at the Jassar Discharge Station using historical streamflow records and hydro-meteorological datasets such as precipitation, temperature, and remotely sensed environmental variables. The intern will work on data preprocessing, exploratory analysis, feature engineering, and implementation of machine learning techniques for predicting streamflow at different forecasting horizons. Various approaches such as regression methods and machine learning algorithms may be explored and compared. The project will provide hands-on experience in hydrological data analysis, predictive modeling, and its applications in water resources while contributing toward improved forecasting capabilities for flood response and water management studies.
Community, Control and Conservation: A Comparative Study of Indigenous Water-Sharing Systems in Chitral and Colonial Irrigation Institutions in Punjab
Mentor: Faris Meher Ali, Communications Specialist at WIT.
Majors: Political Science, History, Anthropology, Law, Public Policy & Sociology.
Necessary Skills: Qualitative analysis, archival research, policy analysis, surveys, interviews, academic writing.
Description: Water governance in Pakistan reflects two contrasting institutional logics: indigenous, community-managed systems in mountainous regions and centralized, colonial-era irrigation structures in the plains. This study compares indigenous water-sharing practices in Chitral with the canal irrigation system in Punjab. While both systems aim to allocate scarce water resources for agriculture, they differ significantly in governance structure, social embeddedness, and equity outcomes. The research investigates how water is distributed, who controls access, and how institutional design shapes justice, efficiency, and sustainability under conditions of scarcity.
The project seeks to compare indigenous water governance systems in Chitral with colonial canal irrigation institutions in Punjab, analyzing decentralized (community-based) and centralized (bureaucratic) mechanisms of water distribution. It further examines socio-political factors influencing water access, including land ownership patterns and class relations, while evaluating equity, efficiency, and sustainability across both systems. Particular attention will be paid to how each system manages conflict, scarcity, and allocation disputes. The study contributes to understanding water not only as a hydrological resource but also as a socially governed system shaped by history, power, and institutional design, providing a basis for rethinking water governance in Pakistan under increasing climate variability and agricultural stress. The expected outcomes of the project include a presentation, a research report, and, subject to the quality and scope of findings, a manuscript suitable for publication.
