Top 10 Companies Hiring AI Engineers in Bangladesh in 2026

Table of Contents

• The “Top 10” Problem (and Why This List Still Matters)
• Grameenphone (GP)
• bKash
• Brain Station 23
• Pathao
• Daraz Bangladesh
• TigerIT
• Chaldal
• DataSoft
• Robi Axiata
• BRAC
• Choosing Your Own “Number 1”
• Frequently Asked Questions

Grameenphone (GP)

Among Bangladesh’s legacy giants, Grameenphone is the clearest case of a telco rebuilding itself as an AI-first company. Its “AI & I” transformation program, covered by The Daily Star’s report on AI & I, explicitly frames GP as an “AI-native” telco-tech firm. For an AI engineer, the draw is simple: models trained on behavior from 80+ million subscribers, the largest consumer data sandbox in Bangladesh.

That scale shows up across core initiatives:

• Churn prediction models that score customer risk at national scale
• Next-best-offer engines personalizing packs and campaigns in near real time
• Self-optimizing networks that tune radio and core parameters based on traffic and outage risk

Technically, you work in a modern MLOps environment rather than a traditional BI shop. GP teams use Python and PyTorch for modeling, with petabyte-scale analytics on Google Cloud Platform and BigQuery. Feature stores, high-volume streaming, and continuous experimentation feed both marketing systems and network operations, so you touch everything from congestion prediction to campaign uplift.

Career-wise, GP runs a centralized Center of Excellence that partners with business units, ideal if you enjoy building reusable AI platforms instead of one-off models. Dhaka-based cross-functional squads mix data engineers, ML engineers, and product owners. Compensation usually tracks the upper mid-range of the local market: mid-level ML engineers often land around 120,000-180,000 BDT/month, while seniors align with roughly 170,000-300,000+ BDT/month, consistent with senior ranges seen in global AI engineer salary guides. If you want to learn how large-scale recommenders and network optimization really work - without leaving Dhaka - GP is one of the strongest possible starting points.

bKash

If Grameenphone owns our call patterns, bKash owns our money flows. As the dominant mobile financial service in Bangladesh, its AI teams sit where risk is real: fraud, credit default, and KYC decisions that can block or enable livelihoods across Dhaka, Chattogram, and beyond.

The bulk of bKash’s applied ML revolves around financial security and intelligence, closely mirroring use cases highlighted in Nucamp’s guide to AI in Bangladesh’s financial services:

• Real-time fraud detection on millions of daily transactions
• Credit scoring models for micro-loans and pay-later products
• NLP-driven support that routes and answers Bangla/English customer queries

On the engineering side, bKash blends mature fintech stacks with production-grade ML: Java/Kotlin for transaction backends, Python for modeling, and AWS plus Kafka for high-throughput streaming. Latency, auditability, and regulatory compliance shape almost every design decision, so you learn to build models that are not just accurate but also explainable and safe enough for risk and compliance teams to sign off.

Compensation reflects its position as a financial heavyweight. According to Glassdoor’s salary data for bKash Dhaka roles, Senior Officer positions pay roughly 714,000-1,000,000 BDT/year (about 60k-83k/month), while Manager roles sit around 1,000,000-2,000,000 BDT/year (about 83k-166k/month). Specialized AI and ML engineers working on fraud or risk typically land at the upper end of these bands or negotiate beyond them. Interviews are known to lean heavily on system design and security-first coding, which makes bKash a strong choice if you want your AI skills tested by both engineering and risk teams.

Brain Station 23

For Dhaka engineers who dream of competing head-to-head with teams in Bengaluru or Singapore, Brain Station 23 feels closest to that world while still being able to commute from Mirpur or Bashundhara. It’s one of the country’s largest software exporters, with a deep bench of AI and data talent serving banks, enterprises, and overseas startups.

Their AI portfolio spans a wide range of “boring but mission-critical” problems that real clients will pay for:

• AI-driven lead scoring for sales organizations struggling with noisy CRMs
• Compliance automation and anomaly detection for banking and financial clients
• Generative AI knowledge bases that let employees query internal documents via LLMs

In a recent showcase of their work on AI-powered digital transformation for operational excellence, Brain Station 23 reported about a 35% reduction in manual errors for some banking clients after deploying ML-powered process automation. Typical stacks combine .NET or Java for core systems with Python-based ML, cloud services (Azure or AWS), and retrieval-augmented generation pipelines for knowledge-heavy use cases.

Because most projects are export-focused, you get practical exposure to GDPR, SOC2, and enterprise-grade MLOps patterns. That means writing models that not only work, but also pass European auditors and large-bank risk committees. Independent rankings of top AI development firms in Bangladesh frequently put Brain Station 23 near the top, with Clutch reviewers giving it a 5/5 rating in AI/ML categories and praising both technical depth and communication.

All of this is reflected in pay. Mid-level AI engineers here typically earn around 120,000-180,000 BDT/month, while senior ML or GenAI engineers often reach the 200,000-250,000+ BDT/month band, in line with what export-oriented firms use to retain talent against remote and regional offers. If you want enterprise AI on global standards without leaving Dhaka, Brain Station 23 is a strong contender to circle on your personal shortlist.

Pathao

Pathao sits where Dhaka feels most alive to a data scientist: ride-sharing, food delivery, and courier services all colliding with chaotic traffic, sudden rain, and hartals. Every surge in Banani or traffic jam in Agrabad turns into numbers your models need to predict.

Logistics analyses like nuVizz’s 2026 AI-in-logistics trends describe this space as the “intelligent backbone” of modern delivery networks. At Pathao, that backbone looks like:

• Dynamic pricing that reacts to demand spikes, weather, and time-of-day
• Demand forecasting for rides and food orders across Dhaka, Chattogram, Sylhet and beyond
• GPS-based route optimization that has to cope with missing signals, roadblocks, and micro-streets

The tech stack blends data science with high-performance backend work: Python for analysis and model training, TensorFlow and sometimes PyTorch for deeper models, and high-concurrency microservices in Go and Node.js. Services are containerized and orchestrated with Kubernetes, so you get hands-on experience pushing low-latency ML APIs to production that must work on low-end Android phones and patchy 3G in Gazipur as reliably as 4G in Gulshan.

Unlike pure analytics roles, feedback loops here are brutally fast: tweak a dispatch model in the morning, and you may see its impact on rider waiting times by evening. That kind of real-time A/B experimentation culture is exactly what logistics-focused consultancies such as Daemon’s engineering case studies point to as critical for high-performing delivery platforms.

On compensation, Pathao’s AI engineers sit in the mainstream Dhaka product-company band. Strong juniors and mid-levels typically earn around 80,000-150,000 BDT/month, while seniors responsible for pricing, dispatch, or ETA algorithms often extend toward 200,000 BDT/month+, aligned with national ML salary snapshots for product-focused roles. If you want to see your model decisions play out on real streets within hours, Pathao offers one of the most intense learning environments in the country.

Daraz Bangladesh

Daraz brings together Alibaba’s AI playbook and the beautiful chaos of Bangladeshi e-commerce: 11.11 and 12.12 mega-campaigns, flash sales that melt servers, and a customer base that ranges from Rangpur students to Gulshan boutique owners. A detailed project report on Daraz Bangladesh’s digital strategy highlights how aggressively the company leans on data and automation to keep that machine running.

At the core is a mix of classic and generative e-commerce AI:

• “AskDaraz”, an AI shopping assistant that lets users search and compare via chat
• Recommendation engines powering homepages, product pages, and email pushes
• Supply-chain and inventory optimization to align regional warehouses with demand spikes
• Dynamic customer segmentation to target offers across hundreds of micro-cohorts

Under the hood, you work with a stack shaped by Alibaba: proprietary Alibaba Cloud analytics and ML platforms, combined with Python-based custom models when local nuance beats global templates. Personalization features are wired directly into modern front-ends, so you see how ranking models, search relevance, and discount engines affect click-through in real time during big campaigns.

Because Daraz operates across South Asia, experimentation culture is strong: teams run continuous A/B tests on layouts, ranking strategies, and incentives. That makes it one of the best local labs for learning data-driven product thinking - why a 0.3% gain in conversion on 11.11 can mean more than a clever new model on a Jupyter notebook.

Compensation tracks the upper end of Dhaka’s product-company range. Solid mid-level AI/ML engineers generally fall around 90,000-180,000 BDT/month, while senior data or ML leads with regional responsibilities often reach the 200,000-250,000 BDT/month band. If your goal is to master large-scale personalization and growth analytics without leaving Bangladesh, Daraz deserves a bright highlighter circle on your Top 10 list.

TigerIT

Hidden behind everyday SIM registrations and e-gates, TigerIT is the Dhaka-born company powering some of the world’s largest digital identity systems. Its biometric platforms, showcased in detail on TigerIT’s biometrics solutions page, sit under national ID and voter databases in multiple countries, often handling millions of biometric records with strict legal and security requirements.

The core AI work revolves around high-accuracy, high-scale recognition problems:

• 1:N biometric matching for national ID and voter systems across tens of millions of citizens
• Face recognition tuned for local lighting, camera quality, and demographic variation
• Iris recognition and secure digital identity verification for border control and law enforcement

According to project summaries on TigerIT’s global projects list and coverage on Biometric Update, their deployments must meet extremely low false-accept and false-reject rates while still performing quickly on-prem, often without cloud. That’s why the stack leans on performance-heavy C++ for matching engines, with Python for research, experimentation, and model training. You’ll live and breathe ROC curves, equal error rates, and template security, not just Kaggle-style accuracy scores.

On the career side, TigerIT is consistently mentioned in “top software companies in Bangladesh” roundups as one of the better payers for deep-tech roles. Mid-level AI or computer-vision engineers typically earn around 100,000-170,000 BDT/month, while seniors who own biometric algorithms or system architecture can reach 200,000 BDT/month+, reflecting the niche, export-grade expertise required.

If your idea of AI is closer to passports and borders than promo banners and click-through rates, TigerIT offers one of the few paths in Bangladesh where your models literally decide identity at national scale.

Chaldal

Chaldal is what happens when the “online bazar” idea grows up into a full-blown AI-powered logistics network. Behind the grocery app that Mirpur or Halishahar families use at midnight sits a dense layer of forecasting, routing, and warehouse automation tuned to Dhaka and Chattogram’s rhythms.

AI is now central to how Chaldal runs its dark warehouses and delivery fleets. In line with global trends described in analyses of AI-driven logistics optimization, the team focuses on:

• SKU-level demand forecasting for each dark store, down to dal vs detergent
• Route optimization for thousands of daily deliveries across dense Dhaka lanes and Chattogram hills
• Computer vision in warehouses for product identification, counting, and basic quality checks

Technically, it’s a full-stack ML environment rather than a pure data-science lab. Python and scikit-learn handle most forecasting and classical ML; PyTorch/TensorFlow power the vision models that read labels and crates on fast-moving belts. Those predictions feed directly into picker apps and rider navigation, so any error shows up immediately as a missing onion in Mohammadpur or a delayed order in Nasirabad.

Targets are unforgiving: similar warehouse-automation systems benchmark 99.9%+ accuracy for stock counting and item recognition, because even a tiny error rate becomes huge at Chaldal’s order volumes. That pressure forces you to care about edge cases, data drift, and monitoring just as much as model architecture.

Compensation is strong for a VC-backed product company. Entry-to-mid AI engineers typically earn around 70,000-130,000 BDT/month, while seniors leading forecasting or routing systems trend toward 150,000-200,000 BDT/month, consistent with export-aware AI salary snapshots from platforms like Remote Recruit’s Bangladesh hiring guide. If you want your models to move physical goods - not just dashboards - Chaldal is one of the sharpest labs in the country.

DataSoft

While many Dhaka engineers chase fintech or ride-sharing, DataSoft has quietly become one of the strongest places to work on Industrial IoT and the “digital RMG factory.” From factories in Dhaka and Narayanganj to clusters in Gazipur and Chattogram, their teams sit at the intersection of AI, manufacturing, and government pushes like Digital Bangladesh and the Hi-Tech Parks that are nudging exporters to automate.

According to their own profile on DataSoft Systems Bangladesh’s LinkedIn, and recent AI/ML infrastructure analyses, the company’s AI work concentrates on:

• AI-driven production planning for RMG lines, predicting machine availability, line efficiency, and on-time delivery risk
• Computer-vision-based quality inspection to cut defect rates and rework before garments leave the floor
• Compliance and audit analytics so suppliers can satisfy demanding EU and US buyers on safety and ESG metrics

The tech stack reflects the reality of Bangladeshi factories: a mix of on-prem servers and cloud (often Azure or AWS) connected to PLCs, IoT gateways, and sensor networks. Python is the core language for ML and anomaly detection on streaming machine data, while dashboards and alerting tools feed supervisors on the line in real time. Instead of click logs, you’re dealing with vibration sensors, energy readings, and machine-status codes.

Industry roundups like Ontik Technology’s list of top software companies frequently cite DataSoft among the country’s leading, better-paying employers. Community salary discussions and roadmap reports suggest mid-level AI engineers typically start around 80,000-120,000 BDT/month, while seniors and solution architects selling to foreign buyers often earn in the 150,000-220,000 BDT/month range. If you want to be the person who explains AI both to a Gazipur factory manager and a Berlin sourcing officer, DataSoft is one of the most practical bridges between code and Bangladesh’s export backbone.

Robi Axiata

Robi Axiata is the other big telco where “data exhaust” has quietly turned into an AI product. Sitting on years of call-detail records, app usage, and network logs from tens of millions of subscribers, Robi’s analytics and AI teams focus less on brand campaigns and more on understanding how customers actually move, call, and churn across Bangladesh.

Industry roundups of top AI players, such as DesignRush’s list of AI development companies in Bangladesh, consistently highlight telecom as one of the most mature verticals for applied ML. Robi’s work reflects that maturity through use cases like:

• Location-based analytics to guide marketing, retail footprint, and urban insights
• Real-time customer-experience monitoring that flags dropped-call hotspots and data-speed anomalies
• Churn prediction and campaign optimization that target specific micro-segments before they port out

Technically, you’re dealing with classic “big data telco” problems. Teams rely on Python or R for modeling, large-scale processing on Hadoop/Spark-style platforms for CDR and network logs, and real-time streaming pipelines feeding NOC dashboards and marketing systems. That means a lot of time-series modeling, anomaly detection, and geo-analytics, often shipped as services that network engineers and business teams use daily.

Career-wise, Robi offers many of the benefits of a large telco without quite the same scale as Grameenphone, which some engineers find more agile. Data/AI analysts and junior ML engineers typically earn around 70,000-120,000 BDT/month, while senior ML engineers or technical leads responsible for churn, experience-scoring, or geo-analytics usually fall in the 150,000-230,000 BDT/month range. Independent ratings on platforms like Clutch’s Bangladesh AI company listings show that telco analytics partners are in high demand globally; Robi gives you a front-row seat to that world from within Bangladesh.

BRAC

For engineers who care more about poverty lines than product funnels, BRAC’s analytics and tech teams offer a very different entry on your Top 10 list. As one of the world’s largest NGOs, BRAC runs programs that touch tens of millions of people across microfinance, health, education, and climate resilience - fertile ground for data-driven decision-making.

Their AI work mirrors the broader “AI for development” themes highlighted in analyses of AI opportunities and challenges in Bangladesh:

• Alternative credit scoring for micro-lending, using non-traditional signals to serve people without formal banking histories
• Predictive health analytics that flag high-risk communities for targeted interventions
• Program evaluation models that sift through survey and outcome data to see which interventions actually work

Technically, this is more “make every taka of donor money count” than “growth hacking.” Teams lean on Python and R, open-source stacks, and a mix of survey data, satellite imagery, and longitudinal program records. Fairness, transparency, and robustness in low-resource settings matter as much as ROC curves: models must be explainable to field officers in Kurigram as well as data scientists in Dhaka.

“The future of work in Bangladesh will not be decided by machines, but by how quickly our people learn to work with them.” - Nur Uddin Ahammed, writing on AI & automation in Bangladesh, LinkedIn

BRAC will not top this list on pay, but it is competitive for the development sector. Practical guides to the local software market, such as Mir Mursalin Ankur’s overview on navigating Bangladesh’s software industry, suggest data/AI roles typically fall in the 60,000-130,000 BDT/month band depending on seniority. The trade-off is clear: you may earn less than at a telco or fintech, but your models can reshape access to credit, healthcare, and education at national scale - turning your loss function into something much closer to lived impact.

Choosing Your Own “Number 1”

Standing back from the notice board, you eventually realise the Top 10 list never knew what you wanted. It just sorted people by its own logic - GPA, quotas, cut-off scores. The same is true for any “Top 10 companies for AI engineers in Bangladesh” ranking: useful as a shortcut, dangerous if you treat it as destiny.

This list was built on four levers - AI maturity, data scale, learning/export exposure, and compensation - but your own objective function may be very different. You might care more about working in Chattogram than Dhaka, or about social impact instead of stock options, or about getting deep into LLM tooling rather than classical ML on tabular data.

A practical way to treat this Top 10 is as a starting map, not a scoreboard. For each company you’ve “circled,” ask yourself:

• Does the domain (telco, fintech, logistics, RMG, NGO) match the problems I want to think about daily?
• Will the stack and data (LLMs, vision, time-series, geo) actually build the skills I want for my next 5 years?
• Is the trade-off between salary, brand name, and ownership of work acceptable for my stage of life?

Then go beyond this article. Check employee stories and client feedback on directories like GoodFirms’ AI company listings for Bangladesh, and compare how different firms are positioning themselves in generative AI on platforms such as RightFirms’ generative AI rankings. Talk to seniors in your university groups, scroll old Facebook threads, and quietly benchmark offers against the ranges you’re seeing in the market.

Once you understand how the list is made - and what it leaves out - you stop asking, “Which company is number one?” and start asking, “Which one is number one for me?” That shift, from rank to fit, is exactly how a good AI engineer thinks: define the objective function first, then optimise. The notice board will always be crowded; the small print is where your real options live.

Frequently Asked Questions