Full-stack engineer building AI, real-time systems & scalable web apps ๐. Turning ideas๐ก โ products ๐
Joined February 2024
- Tweets 288
- Following 23
- Followers 53
- Likes 686
134 Photos and videos
May 22
Day 64/100 of #100DaysOfCode ๐ณ๐
Today I explored Docker Networks & Volumes.
๐ Networks let containers communicate
๐พ Volumes let containers remember data
Realization: Containers alone are temporary.
Networks Volumes make apps scalable & production-ready ๐ฅ
#DevOps
2
80
May 21
Day 63/100 of #100DaysOfCode ๐
Today I finally understood Dockerfiles ๐ณ
A Dockerfile is basically a recipe for creating the SAME app environment anywhere ๐
No more:
โ โWorks on my machineโ
โ Dependency chaos
โ Version mismatch
DevOps journey begins ๐ฅ
#Docker #NodeJS
4
124
May 20
Day 62/100 of #100DaysOfCode ๐
Today I explored Agentic AI ๐ค
AI that doesnโt just respondโฆ
It plans, reasons, uses tools, adapts, remembers context, and gets tasks DONE.
Weโre moving from:
โAI that answersโ
โก๏ธ โAI that actsโ ๐ฅ
#AI #AgenticAI #LLM #Coding
2
1
120
May 12
Day 61/100 of #100DaysOfCode ๐ณ๐ฅ
Today I finally understood Docker Images vs Containers.
๐ Image = Recipe/Blueprint
๐ Container = Running meal/app
One image can create multiple containers ๐
โBuild once. Run anywhere.โ finally makes sense ๐ก
#Docker #DevOps #Coding
3
106
May 11
Day 60/100 of #100DaysOfCode
๐Today I explored TypeScript utility types Zod inference ๐ฅ
โ
Pick
โ
Partial
โ
Readonly
โ
Record
โ
Exclude
โ
Mapped Types
โ
z.infer()
Big realization: Great types donโt just prevent bugsโฆ they design better systems โก
#TypeScript #Zod
1
40
May 10
Day 59 of #100DaysOfCode: Text-to-SQL Generator ๐พ
Logic:
Dynamic Schema Extraction โก๏ธ LLM Translation โก๏ธ SQLite Execution.
Tech: Python, LangChain, Ollama (DeepSeek-R1 local), Streamlit.
Deterministic SQL with temp=0 for accuracy. ๐ฏ
#AI #Python #SQL #OpenSource #GenAI
1
2
141
May 9
๐ Day 58 of #100DaysOfCode
Today I explored AWS โ๏ธ
Learned about EC2, S3, IAM, Route53, Lambda, RDS, DynamoDB, ECS, ECR, Bedrock, SageMaker & more.
Finally understanding how scalable apps are built, deployed, secured, containerized & powered by AI ๐ฅ
#AWS #CloudComputing
3
97
May 7
Day 57 of #100DaysOfCode ๐
Today I explored the real power of TypeScript:
โข Type Inference vs Annotations
โข Interfaces vs Types
โข Generics <T>
โข Union & Intersection Types
Good types donโt just prevent bugs โ they design better systems โก
#TypeScript
3
56
May 6
Day 56/100 ๐
LLMs donโt execute toolsโฆ they suggest them ๐คฏ
Built tool calling in LangChain:
๐งฉ Bind tools
๐ค LLM suggests
โ๏ธ Backend executes
From โjust textโ โ real-world actions ๐ฅ
This is where AI gets serious.
#100DaysOfCode #AI #LangChain #LLM #BuildInPublic
1
5
114
May 5
Day 55/100 ๐
Diving into TypeScript basics & finally getting type safety
๐Arrays โ
Tuples โ
Enums โ
Big one:
โ ๏ธ any = chaos
๐ก๏ธ unknown = safety
๐ฅ Always prefer unknown over any ๐ฅ
JS โ TS = guesswork โ engineering ๐ป
#100DaysOfCode #TypeScript #DevJourney
1
46
May 4
Day 54/100 ๐
Built an AI Image Generation App ๐
Not just prompts โ but a full system:
๐ง Smart prompt engineering
โก FastAPI backend
๐จ Streamlit UI
From text โ ads โ try-ons โ scenes
AI isnโt just tools anymoreโฆ
Itโs experiences. ๐ฅ
#100DaysOfCode #AI #BuildInPublic
1
4
145
May 2
Day 53/100 ๐
Access Token = ๐๏ธ short-lived entry
Refresh Token = ๐ renews access
No localStorage โ (XSS risk)
Secure cookies memory โ
Mind blown ๐คฏ: OTP isnโt storedโonly its hash. Server verifies what it never โknows.โ
Security = layers, not features ๐
#100DaysOfCode
1
2
71
May 1
Day 52/100 ๐
Today I explored Tools in LangChain ๐ง
โ Built-in tools (plug & play)
โ Custom tools (your logic)
โ Creating tools (function description = power)
โ Toolkits (bundled capabilities)
LLMs are no longer just chatbotsโฆ they ACT โก
#100DaysOfCode #AI #LangChain
2
95
Apr 30
Day 51/100 ๐
Learned the backbone of web security:
๐ Hashing Salting (no plain passwords)
๐ก๏ธ bcrypt (slow = secure)
๐ช Cookies flags (HttpOnly, Secure, SameSite)
๐ Tokens for scalable auth
Security isnโt optional. Itโs the foundation.
#100DaysOfCode #WebSecurity #WebDev
2
67
Apr 29
Built a YouTube chatbot with LangChain ๐คฏ
๐ฅ โ Transcript โ Chunks โ Embeddings โ FAISS โ Retrieval โ Prompt โ LLM โ Answer
It doesnโt โwatchโโฆ it understands ๐
RAG in action = smarter, grounded AI ๐
#AI #LangChain #RAG #100DaysOfCode
1
4
160
Apr 28
๐ Day 49 of #100DaysOfCode
Learned the difference between:
๐ Authentication (Who you are)
๐ Authorization (What you can access)
โ
Validation (Is input correct?)
๐ Verification (Is it really you?)
Security = Layered Trust โก
#WebDev #Backend
3
66
Apr 27
Day 48/100 ๐
Finally understood RAG ๐คฏ
Itโs not about making LLMs smarterโฆ
Itโs about giving them the right context at the right time
Index โ Retrieve โ Augment โ Generate
Less hallucination โ
More accuracy โ
Building real AI now ๐ช
#100DaysOfCode #AI #RAG #GenAI
2
60
Apr 26
๐ Day 47 #100DaysOfCode
Revised Prisma ORM, SQL Joins & Foreign Keys ๐ก
INNER, LEFT, RIGHT, FULL JOIN finally make sense ๐ฅ
Big realization: Databases arenโt just storageโฆ theyโre about relationships & connections ๐คฏ
Still learning. Still building ๐
#SQL #Prisma #WebDev
3
107
Apr 25
๐ Day 46/100: Finally understanding how AI really works ๐คฏ
RAG = AI looks up info before answering ๐
Fine-tuning = making AI a specialist ๐ฏ
Embeddings = words โ numbers ๐ข
Big lesson: AI is only as smart as the data we give it.
#100DaysOfCode #AI #MachineLearning #GenAI
3
55
Apr 24
Day 45/100 ๐
Backend concepts finally clicked ๐ฅ
โ Mongoose = structure for MongoDB
โ Relationships (1-1, 1-M, M-M)
โ app.use(express.json()) = req.body works
โ express.static() = serve files easily
Small concepts. Massive impact.
#100DaysOfCode #NodeJS #MongoDB #Backend
4
69