ML engineer, data scientist, AI engineer, research scientist. These four titles overlap a lot and differ in ways nobody explains out loud. You do not have to choose one today, or this year. This lesson is the map: who does what, what they earn, what their day actually looks like, and how AI is reshaping each one.
Learning Objectives
After this lesson, you will be able to:
Identify the eight core roles on a tech team and explain what each one does in plain language
Describe a typical day in the life of each role, including the tools they use daily
Understand salary ranges for each role in India across experience levels
Know how AI is transforming every tech role — and why that is an opportunity, not a threat
Figure out which role might be the best fit for your personality and interests
There is a role for every kind of person. Whether you love art, math, talking to people, solving puzzles, or organizing chaos -- there is a tech role that fits your personality. You do not have to be a "coder" to thrive in tech. This lesson will help you find your fit.
Software does not build itself. Behind every app you use -- Instagram, Zomato, Paytm, Google Maps -- there is a team of people with very different skills working together.
Think of it like making a Bollywood movie. You do not just need actors. You need a director, a cinematographer, set designers, sound engineers, editors, a music composer, and a producer to fund the whole thing. Everyone has a different job, but they all contribute to the final product.
Let us dive deep into each role. For every role, you will learn what they do, what a typical day looks like, what tools they use, what skills they need, how much they earn in India, and how AI is changing their work.
Software engineers write the code that makes apps, websites, and software products work. They are the builders. When a product manager says "we need a search feature," the software engineer is the person who actually builds it.
There are three flavors:
Frontend Developer -- builds what you see. The buttons, the layouts, the animations, the forms. When you scroll through Instagram and the feed loads smoothly, a frontend developer made that happen.
Backend Developer -- builds what you don't see. The server that stores your data, the API that sends your messages, the logic that calculates your Uber fare. When you send a UPI payment and it reaches the right account in 2 seconds, a backend developer made that happen.
Full-stack Developer -- does both. Common in startups where one person needs to handle everything.
A day in the life (9 AM to 6 PM)
Time
Activity
9:00 AM
Check Slack messages, review overnight alerts
9:30 AM
Daily standup -- 15 minutes, each person says what they did yesterday, what they will do today, and any blockers
10:00 AM
Pick up a Jira ticket (task) and start coding
12:00 PM
Lunch break
1:00 PM
Continue coding, write unit tests for the feature
3:00 PM
Code review -- read a teammate's code and give feedback
4:00 PM
Fix bugs found during review, push code to Git
5:00 PM
Deploy the feature to a staging environment and test it
5:30 PM
Update Jira ticket status, write a quick summary in Slack
Backend: Python (FastAPI/Django), Node.js, Java (Spring Boot), Go
Database: PostgreSQL, MongoDB, Redis
Communication: Slack, Jira, Notion
AI Assistants: GitHub Copilot, Claude Code, Cursor AI
Skills required
Technical: At least one programming language deeply (JavaScript/TypeScript or Python), data structures and algorithms, Git, API design, database basics
Soft skills: Problem-solving, ability to break big problems into small ones, clear written communication (code reviews are essentially writing), patience for debugging
Salary range in India (per annum)
Experience
Range (INR)
Fresher (0-1 years)
4-10 LPA
Mid-level (2-5 years)
10-30 LPA
Senior (5-10 years)
25-60 LPA
Staff/Principal (10+ years)
50 LPA - 1 Cr+
Note: Top-tier companies (Google, Microsoft, Uber, etc.) pay significantly higher. Remote roles for US companies can pay 2-3x these ranges.
How AI is changing this role
AI coding assistants like GitHub Copilot and Claude Code can now write boilerplate code, suggest entire functions, and even fix bugs. This does NOT mean developers are being replaced -- it means they are becoming faster. The skill is shifting from "writing every line by hand" to "reviewing AI-generated code, designing systems, and solving the hard problems that AI cannot." Think of it like calculators for mathematicians -- the tool handles arithmetic so you can focus on higher-level thinking.
If data is oil, data engineers build the pipelines that extract, transport, and refine it. They make sure the right data gets to the right place, at the right time, in the right format. Without data engineers, data scientists would have no clean data to analyze and ML models would have nothing to train on.
A day in the life (9 AM to 6 PM)
Time
Activity
9:00 AM
Check monitoring dashboards -- did all overnight pipelines run successfully?
9:30 AM
Investigate a failed pipeline -- a source API changed its format and broke the ETL job
10:30 AM
Fix the pipeline, add better error handling, redeploy
11:00 AM
Meet with the data science team -- they need a new dataset joining three different sources
12:00 PM
Lunch
1:00 PM
Design and build the new data pipeline using Airflow and dbt
3:00 PM
Optimize a slow SQL query that is taking 45 minutes to run on BigQuery
4:30 PM
Write data quality checks to validate that the pipeline output is correct
5:30 PM
Document the new pipeline and update the data catalog
Tools they use
Languages: Python, SQL (the most important skill), Scala
Soft skills: Attention to detail (one wrong join can corrupt millions of rows), patience for debugging data issues at 2 AM, ability to communicate data architecture to non-technical stakeholders
Salary range in India (per annum)
Experience
Range (INR)
Fresher (0-1 years)
5-10 LPA
Mid-level (2-5 years)
12-30 LPA
Senior (5-10 years)
30-55 LPA
Staff/Principal (10+ years)
50 LPA - 1 Cr+
How AI is changing this role
AI tools can now auto-generate SQL queries, suggest pipeline optimizations, and even detect data quality issues automatically. Tools like dbt with AI integration can write transformation logic from natural language descriptions. But data engineering is fundamentally about understanding messy real-world data systems -- something AI struggles with. The role is shifting from "write ETL scripts by hand" to "design data architectures and use AI to accelerate pipeline development."
Data scientists are the detectives of the tech world. They dig through data to find patterns, answer questions, and help the business make smarter decisions. "Which users are likely to cancel their subscription?" "Which product feature drives the most engagement?" "Is this marketing campaign actually working?" -- a data scientist answers these questions with data, not gut feelings.
A day in the life (9 AM to 6 PM)
Time
Activity
9:00 AM
Check experiment results -- did the A/B test for the new checkout flow show a statistically significant improvement?
9:30 AM
Standup with the product team -- share findings and discuss next priorities
10:00 AM
Explore a new dataset -- use Pandas and Matplotlib to understand distributions, outliers, and patterns
11:30 AM
Clean the data -- handle missing values, remove duplicates, engineer new features
12:00 PM
Lunch
1:00 PM
Build a churn prediction model using Scikit-learn -- try different algorithms, tune hyperparameters
3:00 PM
Evaluate model performance -- precision, recall, ROC curve analysis
4:00 PM
Create a presentation for the business team explaining the findings in non-technical language
5:00 PM
Peer review a colleague's analysis
Tools they use
Languages: Python (primary), R (some teams), SQL
Notebooks: Jupyter Notebook, Google Colab
Data manipulation: Pandas, NumPy
ML libraries: Scikit-learn, XGBoost, LightGBM
Visualization: Matplotlib, Seaborn, Plotly, Tableau, Power BI
Communication: PowerPoint / Google Slides (yes, really -- presenting findings is half the job)
Skills required
Technical: Statistics and probability (hypothesis testing, regression), Python, SQL, machine learning algorithms, data visualization, experimental design (A/B testing)
Soft skills: Storytelling with data (the best data scientist is the one who can explain a complex finding to a CEO in 2 minutes), curiosity, business acumen (understanding what questions actually matter)
Salary range in India (per annum)
Experience
Range (INR)
Fresher (0-1 years)
5-12 LPA
Mid-level (2-5 years)
12-35 LPA
Senior (5-10 years)
30-60 LPA
Staff/Principal (10+ years)
55 LPA - 1.2 Cr+
How AI is changing this role
This is the role most directly affected by AI. Tools like ChatGPT can now write exploratory data analysis code, generate visualizations, and even build basic ML models from natural language prompts. The routine parts of data science (cleaning data, running standard models) are being automated. But the strategic parts -- figuring out what questions to ask, designing experiments, and translating insights into business decisions -- still require a human who understands both the data and the business context. The role is evolving from "person who writes Python to analyze data" to "person who uses AI tools to analyze data 10x faster and focuses on strategy."
If a data scientist builds a model that works in a Jupyter notebook, the ML engineer is the person who makes it work in the real world -- handling millions of requests per second, running on cloud servers, and not crashing at 3 AM. They bridge the gap between research and production. They also increasingly build AI-powered features using large language models, vector databases, and agent frameworks.
A day in the life (9 AM to 6 PM)
Time
Activity
9:00 AM
Check the dashboards: is the recommendation model getting worse over time (this is called drift), and is it still answering fast enough to meet the speed promised to other teams (an SLA)?
9:30 AM
Standup -- discuss the new AI-powered feature the team is building
10:00 AM
Take a general-purpose language model and train it a little further on the company's own text so it picks up the house vocabulary. This is called fine-tuning.
12:00 PM
Lunch
1:00 PM
Make the model answer faster. Usually this means converting it to a leaner format and measuring the difference.
3:00 PM
Connect the model to the company's own documents so it can answer questions about them instead of guessing. This pattern is called RAG, and it has a whole track to itself later.
4:30 PM
Write tests that check the model's service still works when other systems call it.
5:30 PM
Read a teammate's proposed code change and leave comments.
Tools they use
ML frameworks: PyTorch (dominant), TensorFlow, JAX
LLM tools: LangChain, LlamaIndex, Hugging Face Transformers, vLLM
Experiment tracking: MLflow, Weights & Biases
Model serving: TorchServe, Triton Inference Server, BentoML
Vector databases: ChromaDB, Pinecone, Weaviate
Infrastructure: Docker, Kubernetes, AWS SageMaker, GCP Vertex AI
Languages: Python (primary), C++ (for optimization), Rust (emerging)
Skills required
Technical: Deep understanding of neural networks and training, Python, PyTorch, Docker/Kubernetes, cloud platforms, model optimization (quantization, distillation, pruning), understanding of LLMs and prompt engineering
Soft skills: Ability to translate research papers into production code, comfort with ambiguity (ML is inherently experimental), strong debugging skills (ML bugs are silent -- the model just performs badly)
Salary range in India (per annum)
Experience
Range (INR)
Fresher (0-1 years)
6-15 LPA
Mid-level (2-5 years)
15-40 LPA
Senior (5-10 years)
35-70 LPA
Staff/Principal (10+ years)
60 LPA - 1.5 Cr+
ML engineers are among the highest-paid roles in tech due to the specialized skill set and high demand.
How AI is changing this role
Paradoxically, AI is making the ML engineer's job both easier and more complex. Tools like AutoML can handle routine model selection. LLM APIs (Claude, GPT-4) mean you do not always need to train a model from scratch -- you can build powerful applications by orchestrating pre-trained models. But someone still needs to design the architecture, handle edge cases, optimize for cost and latency, and make sure the system is reliable. The role is expanding from "train and deploy models" to "architect AI systems using a mix of pre-trained models, fine-tuning, and custom training."
DevOps (Development Operations) and SRE (Site Reliability Engineering) engineers are the people who keep your apps running. When you open Swiggy and it loads in under a second, that is because a DevOps engineer set up the infrastructure. When Netflix handles 200 million users streaming simultaneously without crashing, that is SRE at work. They automate deployments, manage cloud infrastructure, and wake up at 3 AM when things break.
A day in the life (9 AM to 6 PM)
Time
Activity
9:00 AM
Check Grafana dashboards -- CPU usage, memory, error rates, latency across all services
9:30 AM
Investigate an alert -- one microservice's error rate spiked overnight. Check logs in Datadog
10:30 AM
Root cause found: a new deployment had a memory leak. Roll it back
11:00 AM
Write a Terraform script to provision a new Kubernetes cluster for the ML team
12:00 PM
Lunch
1:00 PM
Improve the CI/CD pipeline -- add a new test stage so broken code cannot reach production
3:00 PM
Security patching -- update base Docker images to fix a CVE vulnerability
4:00 PM
On-call handoff meeting -- brief the next person on any ongoing incidents
5:00 PM
Write a runbook for a common incident type so it can be resolved faster next time
Tools they use
Containers: Docker, Kubernetes (the most important skill in modern DevOps)
Technical: Linux administration, networking (TCP/IP, DNS, load balancers), Docker and Kubernetes, at least one cloud platform deeply, CI/CD pipeline design, scripting (Bash + Python), understanding of security principles
Soft skills: Calm under pressure (you will get paged during outages), excellent documentation skills, systematic problem-solving, communication during incidents ("War room" coordination)
Salary range in India (per annum)
Experience
Range (INR)
Fresher (0-1 years)
4-10 LPA
Mid-level (2-5 years)
10-30 LPA
Senior (5-10 years)
25-55 LPA
Staff/Principal (10+ years)
50 LPA - 1 Cr+
How AI is changing this role
AI is helping DevOps engineers by providing intelligent alerting (predicting incidents before they happen), auto-remediation (automatically fixing known issues), and AI-powered log analysis (finding the needle in a haystack of millions of log lines). Tools like Kubernetes copilots can generate YAML configurations from natural language. But infrastructure is fundamentally about reliability and security -- areas where you cannot afford AI hallucinations. The human remains critical for architectural decisions, security posture, and incident response.
The Product Manager (PM) decides what to build and why. They do not write code. Instead, they talk to users, analyze data, understand the market, and create a prioritized plan for what the engineering team should build next. They are the voice of the user inside the company. A great PM can turn a vague business goal ("increase revenue") into a concrete feature spec ("add a subscription tier with these 3 features at this price point").
A day in the life (9 AM to 6 PM)
Time
Activity
9:00 AM
Review overnight user feedback from App Store reviews and support tickets
9:30 AM
Standup with the engineering team -- any blockers? Any scope questions?
10:00 AM
User interview -- hop on a 30-minute call with a customer to understand their pain points
Communication: Slack, Zoom, Loom (async video updates)
Roadmapping: Productboard, Aha!, Notion
User research: Hotjar (heatmaps), UserTesting, Typeform (surveys)
Skills required
Technical: You do not need to code, but you need to understand how software is built (APIs, databases, trade-offs between speed and quality). Enough technical literacy to have informed conversations with engineers
Analytical: SQL (basic to intermediate), understanding of metrics (DAU, MAU, retention, LTV, CAC), A/B testing methodology
Soft skills: Communication (you are the glue between engineering, design, sales, and leadership), prioritization (saying "no" to 90% of ideas), empathy (deeply understanding user pain), storytelling (getting buy-in for your vision)
Salary range in India (per annum)
Experience
Range (INR)
Fresher / APM (0-2 years)
8-15 LPA
Mid-level PM (2-5 years)
15-35 LPA
Senior PM (5-10 years)
30-60 LPA
Director/VP Product (10+ years)
55 LPA - 1.5 Cr+
How AI is changing this role
AI is becoming the PM's superpower. PMs now use AI to summarize user feedback (reading 10,000 reviews in seconds), generate PRDs from rough notes, analyze metrics using natural language queries, and even prototype features with AI coding tools. The role is shifting from "spend 60% of your time on documentation" to "spend 60% of your time talking to users and thinking strategically, while AI handles the paperwork."
QA (Quality Assurance) engineers make sure the software actually works before it reaches users. They find bugs, write automated tests, and prevent broken features from shipping. Think of them as the last line of defense between buggy code and your users. Every time you use an app and it does NOT crash -- thank a QA engineer.
A day in the life (9 AM to 6 PM)
Time
Activity
9:00 AM
Review the latest build -- what new features were deployed to the staging environment overnight?
9:30 AM
Standup -- hear from developers what changed and plan testing accordingly
10:00 AM
Write automated test cases for the new payment flow using Playwright
12:00 PM
Lunch
1:00 PM
Run the full regression test suite -- make sure old features still work after the new code
2:00 PM
Exploratory testing -- manually try edge cases that automated tests might miss ("What happens if I enter emoji in the phone number field?")
3:30 PM
File detailed bug reports in Jira with steps to reproduce, screenshots, and expected vs actual behavior
4:30 PM
Test API endpoints using Postman -- verify response codes, error handling, and edge cases
5:30 PM
Update test documentation and coverage reports
Tools they use
Test automation: Playwright (modern favorite), Cypress, Selenium
API testing: Postman, Bruno, REST Client
Bug tracking: Jira, Linear, GitHub Issues
Performance testing: k6, JMeter, Locust
Mobile testing: Appium, BrowserStack, Sauce Labs
Languages: JavaScript/TypeScript (for test scripts), Python
CI integration: GitHub Actions (running tests on every commit)
Skills required
Technical: At least one programming language (JavaScript or Python), test automation frameworks, API testing, understanding of CI/CD pipelines, basic SQL (for verifying data in the database)
Soft skills: Extreme attention to detail, adversarial thinking ("How can I break this?"), clear communication (writing bug reports that developers can actually reproduce), patience, systematic approach
Salary range in India (per annum)
Experience
Range (INR)
Fresher (0-1 years)
3-8 LPA
Mid-level (2-5 years)
8-22 LPA
Senior (5-10 years)
20-45 LPA
Lead/Architect (10+ years)
40-80 LPA
How AI is changing this role
AI is transforming QA from primarily manual testing to intelligent test automation. AI tools can now auto-generate test cases from requirement documents, identify which tests to run based on code changes (intelligent test selection), and even perform visual regression testing by comparing screenshots. Self-healing tests (tests that automatically update when the UI changes) are becoming reality. The role is evolving from "write and run tests" to "design test strategies and supervise AI-powered testing systems."
UI (User Interface) designers make the product look good. UX (User Experience) designers make the product feel good -- easy to use, intuitive, and delightful. In practice, most designers do both. They decide where every button goes, what color it should be, what happens when you tap it, and how the entire flow feels from start to finish. If a software engineer builds a house, the designer is the architect who plans the floor layout and picks the paint colors.
A day in the life (9 AM to 6 PM)
Time
Activity
9:00 AM
Review analytics from Hotjar -- where are users getting stuck? Where are they rage-clicking?
9:30 AM
User research session -- watch a recording of a user trying to complete a task and note where they struggle
10:30 AM
Sketch wireframes on paper or in Figma -- low-fidelity layouts for a new feature
11:30 AM
Design review with the PM -- discuss the wireframes, align on scope
12:00 PM
Lunch
1:00 PM
Create high-fidelity mockups in Figma -- pixel-perfect designs with real content, colors, and typography
3:00 PM
Build an interactive prototype in Figma -- link screens together so stakeholders can "click through" the flow
4:00 PM
Usability testing -- have 3 people try the prototype while you observe and take notes
5:00 PM
Iterate on designs based on usability feedback
5:30 PM
Hand off final designs to the frontend developer with specs and assets
AI tools: Midjourney/DALL-E (for concept art), Galileo AI, Uizard
Skills required
Technical: Figma proficiency, understanding of design systems, basic knowledge of HTML/CSS (to design within technical constraints), responsive design principles, accessibility (WCAG standards)
Creative: Visual design (typography, color theory, layout), interaction design (micro-animations, transitions), information architecture
Soft skills: Empathy (the whole job is understanding how humans think and feel), ability to take criticism on designs, communication (defending design decisions with data), user research skills
Salary range in India (per annum)
Experience
Range (INR)
Fresher (0-1 years)
3-8 LPA
Mid-level (2-5 years)
8-25 LPA
Senior (5-10 years)
22-50 LPA
Lead/Director (10+ years)
45-90 LPA
How AI is changing this role
AI design tools can now generate UI layouts from text prompts, create variations of designs instantly, and even convert wireframes to code. Tools like Galileo AI and Uizard can produce initial design mockups in seconds. But design is fundamentally about understanding human psychology, business goals, and crafting experiences that feel right -- not just look right. AI accelerates the execution but cannot replace the thinking. The role is shifting from "spend 3 hours creating 5 layout variations" to "use AI to generate 50 variations in 5 minutes, then pick and refine the best one."
This is for people who love doing the work -- writing code, designing systems, analyzing data. You get promoted by becoming more skilled and taking on harder problems:
Fresher (0-1 year) -- You are learning. Assigned simple tasks. Lots of hand-holding from seniors.
Junior (1-2 years) -- You can handle straightforward tasks independently. You still need guidance on bigger decisions.
Mid-level (2-5 years) -- You own features end-to-end. You can design solutions, not just implement them. You mentor juniors.
Senior (5-8 years) -- You are the go-to person on your team. You make architectural decisions. You unblock others. You see the big picture.
Staff (8-12 years) -- You influence the technical direction of multiple teams. You solve problems that nobody else can define yet.
Principal (12+ years) -- You shape the technical strategy of the entire company. You are consulted on the hardest problems. Very few people reach this level.
There is no right answer here, and nothing is being marked. Pick the one that sounds most like you.
Based on what excites you most, which tech role might be the best fit for you?
There is no wrong answer. And here is a secret: most people switch roles at least once in their career. Many great PMs were former engineers. Many ML engineers started as data scientists. Many DevOps engineers started as backend developers.
Here is a quick guide based on your personality:
"I love making things look beautiful and intuitive"
Start with UI/UX Design or Frontend Development. If you lean creative, go design. If you lean technical, go frontend. Many people do both.
"I love solving puzzles with data and finding hidden patterns"
Start with Data Science. Learn Python, statistics, and SQL. If you enjoy the infrastructure side more, pivot to Data Engineering.
"I love building things that work at scale"
Start with Backend Development. If you enjoy the infrastructure and reliability side, move into DevOps/SRE. If you enjoy the ML side, move into ML Engineering.
"I love automating everything"DevOps is your calling. You will write scripts that save hundreds of hours, build pipelines that deploy code automatically, and infrastructure that scales itself.
"I love the cutting edge of AI"AI/ML Engineering is where the action is. Learn Python, PyTorch, and how neural networks work. This platform -- RugvAI Labs -- will teach you everything you need to get started.
"I love understanding users and making strategic decisions"Product Management might be your path. You do not need to code, but you need to understand technology, users, and business deeply.
"I love breaking things and finding bugs"QA Engineering values your mindset. Learn test automation and you will be in high demand.
#The Title Overlap Problem (And How to Pick the Right One)
These four titles confuse everyone:
Data Scientist. Answers business questions with data. Builds models in notebooks. Presents findings. Lives in Jupyter + SQL + Pandas.
ML Engineer. Takes a data scientist's notebook and turns it into a service handling 10M requests per day. Lives in PyTorch + Docker + Kubernetes.
AI Engineer / LLM Engineer. A newer title (2023+). Builds applications using foundation models (Claude, GPT, Gemini). Lives in LangChain + vector databases + prompt engineering. Less ML theory, more API integration.
Research Scientist. Invents new model architectures, publishes papers. Usually requires a PhD. Lives at Anthropic, DeepMind, FAIR. Highest salary ceiling, hardest to enter.
Same person, same skills, four very different titles. Here's how to pick:
If you love...
Aim for
Asking "what does the data say?"
Data Scientist
Taking a model from notebook to production
ML Engineer
Building chatbots, RAG systems, agents
AI Engineer
Reading papers, inventing new algorithms
Research Scientist
All four
Pick one for your first 2-3 years, then pivot — you can't be excellent at all four simultaneously
The titles also signal seniority differently. A "Senior Data Scientist" at a mid-size company often makes less than a "Junior ML Engineer" at a frontier lab. Always research the company-specific level for the title you're interviewing for.
A new role that barely existed in 2022: AI Product Manager. This is a PM who specifically owns AI-powered features. Why is this its own role? Because AI products break in ways traditional products don't:
Non-determinism: the same input can give different outputs. Traditional PMs aren't trained for this.
Evaluation: how do you A/B test "the model writes better summaries"? You need eval frameworks, not just funnel metrics.
Hallucinations: when do you trust the model? AI PMs design the guardrails.
Cost economics: every call to an LLM costs money. PMs must understand unit economics in a way traditional PMs do not.
If you're a current PM, learning the AI side is the highest-ROI move you can make in 2026.
#What Each Role's First 90 Days Actually Looks Like
Recruiters never tell you this. Here's the truth.
Software Engineer (Frontend or Backend) — Day 1-90:
Week 1: get your machine set up. This takes longer than you'd think; expect 2-3 days.
Week 2: read existing code, ask a senior to walk you through the architecture. Don't write production code yet.
Week 3-4: pick up a "starter ticket" — a small bug or minor feature. Your first PR.
Month 2: own a small feature end-to-end. Get used to code review feedback (it's brutal, but it's mentorship, not personal attack).
Month 3: take part in on-call rotation (in some teams).
Data Scientist — Day 1-90:
Week 1-2: understand the data. Where it lives, how it's collected, what's clean, what's broken. This is 80% of the job.
Week 3-4: replicate an existing dashboard or analysis. Validate you can produce the same numbers as the previous person.
Month 2: design and run your first A/B test, or build your first model. Both intimidating; both done in iteration.
Month 3: present findings to stakeholders. The hardest skill: making a complex analysis land with a non-technical executive in 2 minutes.
ML Engineer — Day 1-90:
Week 1: read code, understand the current training pipeline and serving infrastructure.
Week 2-4: optimize something small — reduce inference latency by 10%, fix a memory leak, improve a CI test.
Month 2: own a model in production. Watch it 24/7 for a week — drift, errors, edge cases.
Month 3: ship a new model version with measurable improvements.
Product Manager — Day 1-90:
Week 1-2: 30+ user interviews. Yes, really. You can't PM what you haven't talked to users about.
Week 3-4: shadow the engineering team. Sit in standups, design reviews, sprint planning.
Month 2: own a small feature spec. Get it past design review and into a sprint.
Month 3: own a roadmap item that lands in front of the CEO.
DevOps / SRE — Day 1-90:
Week 1: get on-call shadow rotation. Watch the senior on-call handle real incidents.
Week 2-4: own one piece of infrastructure (a CI/CD pipeline, a monitoring dashboard, a Kubernetes namespace).
Month 2: take your first primary on-call shift. Yes, you'll get paged at 3 AM. Yes, you'll survive.
Month 3: lead a post-mortem after an incident. The post-mortem (not the incident) is where careers are made.
No matter which role you choose, certain skills will make you successful everywhere in tech:
1. Communication
The best engineer in the world is useless if they cannot explain their code to a teammate. The best data scientist is useless if they cannot present their findings to the CEO. In tech, writing clearly -- in Slack messages, code comments, documentation, and emails -- is as important as writing code.
2. Problem-solving
Tech is fundamentally about solving problems. The specific tools change every 3-5 years, but the ability to break a complex problem into smaller pieces, research solutions, and iterate -- that never goes out of style.
3. Continuous learning
The tech industry moves faster than any other. The tools and frameworks you learn today might be obsolete in 5 years. The people who thrive are the ones who are always learning -- reading documentation, taking courses, experimenting with new technologies, and staying curious.
4. Teamwork and collaboration
Software is built by teams, not individuals. Every day involves code reviews, design discussions, standup meetings, and working through disagreements. Being a good teammate -- reliable, respectful, and helpful -- matters more than being the smartest person in the room.
5. English proficiency
This might seem odd in a tech list, but most of the tech world operates in English. Documentation, Stack Overflow, GitHub, Slack channels, meetings with global teams -- all in English. Investing in your English communication skills has one of the highest ROIs in tech.
6. Basic AI literacy
Regardless of your role, understanding the basics of AI -- what it can do, what it cannot do, and how to use AI tools effectively -- is becoming a requirement. This is not optional anymore. Every role in the list above now uses AI tools daily. The people who learn to work with AI will outperform those who ignore it.
Building software is a team sport with distinct roles -- developers build it, designers make it usable, PMs decide what to build, QA makes sure it works, DevOps keeps it running, and data professionals make it smart
Each role has a clear day-to-day rhythm -- from standups and Jira tickets to Figma mockups and Grafana dashboards. No two roles look the same on a daily basis
Career growth follows two tracks -- the IC (Individual Contributor) track for people who love doing the work, and the Management track for people who love leading teams. Both can reach senior levels
T-shaped skills win -- go deep in one area, then broaden. The most valuable professionals understand their specialty deeply AND can collaborate across disciplines
AI is transforming every role -- not replacing them. It is making routine tasks faster and shifting focus to higher-level thinking, strategy, and creativity. Learning to work with AI tools is now a core skill for every tech professional
Communication and continuous learning are universal -- no matter your role, the ability to explain complex ideas simply and the habit of always learning new things will set you apart
The Software Industry -- understand the business context these roles operate in
How Software Gets Built -- see how these roles collaborate through the software development lifecycle
AI Solving Real Problems -- discover the real-world AI applications these roles are building
Quick Check1 / 5
The website is slow. Users are complaining on Twitter that pages take 8 seconds to load. Who investigates first?
Each role you read about above maps to a learning path on this platform. Use these forward-links to jump straight to the technical foundations for the seat you want.
Next: How Software Gets Built. You know who is on the team. The next lesson is how they actually work together: sprints, tickets, code review, and the pipeline that takes a change from a laptop to millions of phones. You have already met the companies these roles work at, the six revenue models, the funding ladder from seed to IPO, and India's evolving role in global tech. Picking the right company is as important as picking the right role.