Hi Friends,
Hemen Parekh
27 June 2013
Now as I approach my 90th birthday ( 27 June 2023 ) , I invite you to visit my Digital Avatar ( www.hemenparekh.ai ) – and continue chatting with me , even when I am no more here physically
Monday, 22 October 2012
Monday, 10 September 2012
VISUAL REPRESENTATION
Enclosed graphs prove my point.
On our "Job Suggest" page, we have counters:
-
Live Jobs
-
Archive Jobs
As soon as Rahul comes out of priority tasks, I would like to replace these counters with attached graphs.
May be, we can just leave the counters as these are today, but make these “clickable.”
Upon clicking, relevant chart will pop-up.
I suppose we are currently maintaining daily logs of both counters, so that, whenever we decide to introduce graphs, there will be no problem.
Monday, 27 August 2012
JOBS APPLIED HISTORY
This is further to my earlier notes/emails on the subject, resting with my last note dt. 24th Aug. 2012.
Pl. let me know, what data (about each applied job) that you plan to capture. I suppose, this could well be, ALL the fields of the job applied—
Even though, at this moment, we may not know, how exactly we are going to use every captured “field”, for recommending more relevant jobs.
But it is better to capture everything. You never know, how it could help in future.
Page 2/2
This is just my preliminary concept and I invite you to suggest changes.
During my study, I found that “Design Positions/Titles” vary widely with small changes. There is no STANDARDIZATION / NORMALIZATION. Hence, computing their “Frequency of Occurrence” and tabulating in descending order, is impossible.
Once you have had an opportunity to study enclosed pages, pl. let me know. We may need to sit together, in order to finalize this matter, BEFORE you start capturing data on 10 Sept.
| Full Advt | In Widget |
|---|
-
Advt ID | | ✓
-
Designation (Position/Title) | | ✓
-
Company Name | | ✓
-
Job Description
-
Desired Profile
-
Compensation
-
Experience (No. of years) | | ✓
-
Industry Type
-
Education | | ✓
-
Location (City) | | ✓
-
Keywords
-
Post Date
-
Expiry Date
Typical “Brief Resume” for an “Experienced Buff” contains following fields
A. Personal Details
-
Name
-
D.O.B.
-
Gender
-
Marital Status
-
Email ID
-
Mobile No.
-
Country
-
City
B. Professional Details
-
Type – ○ IT ○ Non-IT
-
Primary Function
-
Exp. (yrs)
-
Designation Level
-
Current Industry
C. Text Resume
Shuklendu
24-08-12
Jobs Applied History → Job Recommendation System
What might database of "Jobs Applied History" of a candidate, might look like?
PEN: ___________
Name: ___________
| Sr. No | Date Appl. | Time Applied | Online | Offline (Widget) | Job ID | Field Details of Job Advt |
|---|---|---|---|---|---|---|
| Advt ID |
Field No: 11 is "Keywords".
-
A second table will keep “adding” these keywords and
-
Calculate their Frequency of Occurrence and
-
Arrange these keywords in Descending Order of the Frequency of Occurrence and
-
Calculate the Weightages of each keyword (taking top 100 keywords = 100)
As far as "Recommended / Matched Jobs" is concerned, current system of “matching” will continue, till a given candidate has applied against 100 job advts. (– assuming that, by that time, second table has a fairly respectable no. of "Keywords").
Would you like me to extract all 7 pages of this "Jobs Applied History → Job Recommendation System" series into one continuous Word document for you?
How can we leverage “Job Applied History” of each candidate?
-
In how many different “ways”, can we analyze this history?
-
What can such “analysis” tell us? Tell concerned jobseeker? Tell Corp. employers?
-
Will/can such “analysis” answer following questions?
-
How often does each user “log-in” thru widget?
(frequency/day/month/year – Ave/Max) -
How many job-alerts does he view/pull each time?
-
How many jobs does he “apply” each time?
-
Ratio = 3 / 2 =
No. ViewedNo. Applied
{ All of these data must be compiled,
datewise/monthwise/yearwise
– for the day or the month or the year
– Cumulative. }
Page 4/7
-
At what “time-of-the-day” does each jobseeker log-in?
Is there a pattern?
Does each have a “favorite” time? -
During the course of the day, what % of widget-users log-in, at diff. times?
(Load on server) -
✅ Does each candidate have some favorite
-
Companies
-
Cities
to whom he frequently “applies”?
-
On monthly/annual basis, compile, in descending order (of frequencies of “Apply”):
-
Companies to which he applied
-
Cities to which he applied
This analysis will reveal what are his favorite Companies / Cities.
Knowing his “favorites” will help us recommend to him, future jobs advertised by
-
same companies
-
in same cities.
(Database of Intentions?)
Once a person takes up a job after graduation, his "Edu. Qualification" changes only in rare cases (maybe 5%).
So, it is NOT important to compile “jobs applied history” for jobs demanding:
-
B.A. or
-
B.Sc. or
-
M.B.A.
→ a specific Edu. Quali.
But, over the years, a person's experience grows in terms of “No. of years”.
And his job preferences will also change,
from
0–2 years → when he is FRESH
2–4 " → " " " JUNIOR
6–8 " → " " " Sup.
8–10 " → " " " Manager
10–12 " → " " " Sr. Mgr
12–15 " → " " " G.M. etc.
So, over the years, you will find him applying for jobs with increasing years of exp. (required)
This will be a significant trend.
Now correlate this with his "age" (derived from his birthdate and adding one year to his age, on each birthday)
(Graph illustration)
Y-axis: "Jobs demanding Total Exp. (yrs)"
X-axis: "Age"
Range:
-
X-axis: 20 to 45
-
Y-axis: 2 to 14 (years of experience)
A diagonal trend line shows an increase in experience demand with age.
From "History of Jobs Applied", we can compile such “correlation tabulation” (if not a graph), for:
-
each candidate
-
all the candidates
That would tell us a lot about
“What” kind (how many years of demanded exp.)
of jobs, should we recommend to
“What” kind of candidates (how old candidates)
So, in final, we must analyse,
“Jobs Applied History”
for [ Individuals + All candidates put together ]
A. ✔ No. of Jobs Applied
B. ✔ Companies
C. ✔ Cities
D. ✘ Designations
E. ✘ Exp. (Demanded)
F. ✔ Keywords
Display?
⬜ Year ▼ ⬜ Mon ▼ ⬜ Date ▼
Let us discuss this after you study it.
👇
27–08–12
Tuesday, 21 August 2012
Friday, 10 August 2012
PROJECT DIGITIZATION
Background
-
Over the years, I have sent out handwritten notes to colleagues
-
Total no. of pages could be 5000+
-
It is occupying lot of space/shelf
Idea
-
Digitize (scan) these pages & prepare soft copies
-
Destroy hard copies
Problem
-
Having created soft copies, how do we “search / find” any particular note in future, if required?
Solution
-
To each scanned copy (Image Page), attach (create database):▶ Date (DD/MM/YY)▶ Person to whom addressed▶ Subject / Title▶ Keywords (found in Image page)
Example of Data to be recorded for each Image
-
▶ Date
-
▶ Person Name (Sometimes “Consultants”)
-
▶ Subject ( “ ” missing)
-
▶ Keywords
HCP will attach such a sheet to each note, before Tushar scans it and enters data via Datasheet.
Also, datasheet to carry:
-
▶ File Name / Folder Name / File Path
What would OFFLINE search page look like?
→ See attached
Image 1: Broad Perspective (Page 4/5)
Broad Perspective
-
To begin with, searching of documents will be OFFLINE, by our consultants.
-
But later on, it should be possible to migrate the entire database (SQL?) and the Search Feature, ONLINE, onCustomizeResume.com
-
To be developed in ASP.net (v4.0)?
-
SQL database (of images/keywords)
-
When uploaded ONLINE, Google spider must be able to click and reach each & every IMAGE and be able to capture that page as an IMAGE and index that page against all the keywords/Metatags found and store it in memory, and deliver that IMAGE as a search-result, when anyone uses any of the keywords as a “Search Term”.
Image 2: Functional Specifications (Page 5/5)
FUNCTIONAL SPECIFICATIONS
IMPORTANT
▶︎ There must be no need for a visitor to “Type / Enter” anything at all and anywhere
▶︎ Make ALL pages “Reachable” by Google Spider
▶︎ No login of any kind. Any visitor can access the database
Saturday, 30 June 2012
JOB POST WIDGET
Background / Assumptions:
▶ For jobs database today, we are dangerously dependent upon RSS feeds from Naukri / TimesJobs / ClickJobs.
▶ We will reduce (but not eliminate) this dependence by getting RSS feeds from MonsterIndia / JobsAhead / Jobstreet and other job portals.
▶ When more Employers register on our website, hopefully, they will post jobs online.
▶ But, in India, perhaps no more than 80,000 employers are using job portals for online recruitment (See the list in our RESUMEBLAST database).
These are employers who have “access” to internet in their offices AND who have/are already using Naukri / Monster / TimesJobs etc.
Page 2/2
▶ But, there are millions of other SMALL / MEDIUM size employers who have NO access to internet and therefore, have never used a jobsite for their recruitment needs.
These are small/medium businesses, retailers, shops, service-repair companies, etc. etc.
▶ Apart from the fact that they have NO PC/Broadband in their offices (to access jobsites), they cannot afford to pay Rs. 25,000/- for a Quarterly Subscription to Naukri etc. just to advertise 1 vacancy/year!
But that does not mean they would not use internet-based recruitment if
-
they can post their job-vacancy (on CustomizeResume, of course), using a Smartphone-based "App"
-
they don't need to pay anything for such posting (FREE)
-
they can receive "Brief Resume" as email on their smartphone
-
If they can click/touch the mobile no. (of applying candidate) directly in that email
In fact, a much better / instant "service" could be:
-
SME employer posts his vacancy (on CustomizeResume) using his smartphone(He would need PC/Internet ONCE in the beginning, to register as Employer and get his User ID / Password, in order to use "Post Job" widget).
| Srl No | Candidate Name | City (Where living) | Edu. Quali | Exp. | DOB | Mobile No |
|---|---|---|---|---|---|---|
| 1 | Jayant Shinde | Sangli | BSc | 4 | 20/1/84 | 98213 46897 |
One way to do this (make money) is by revealing only 5 resume highlights to the employer, at-a-time.
-
If he wants to see more, then, he will need to touch/view a product advt. delivered by Yahoo/InMobi advt. network.
-
What would various screens to "POST JOB" look like?
As usual:
+--------------------------+
| RecruitGuru |
+--------------------------+
| Post Job |
| (Button) |
+--------------------------+
| To use this free "PostJob"
| app – and to receive matching
| resume highlights on your
| mobile, please register
| first (as an Employer) on
| www.Cust...
| and get your UserID / Passwd.
| If registered, click to proceed
+--------------------------+
+-------------------------------+| Select any one || [x] 10th Standard || [ ] 12th Standard || [ ] Diploma || [ ] BA || [ ] B.Com || [ ] B.Sc. |+-------------------------------+
These are my preliminary thoughts
on the subject of
"Post Job" Widget.
Just as we are trying to make job search OBSOLETE, thru MyJobs, idea is to make resume search obsolete thru POST JOBS.
People don't want to spend time "searching".
They have just enough time to "find" things which are relevant to their needs.
Which employer has time/luxury to post his job advt
and then,
wait and wait and wait for the right candidates to "discover" that advt (from lakhs & lakhs) and then "apply"?
Only to discover after 30 days, that
▶ only the WRONG candidates had time to discover your ad and apply!
Big Companies have full-fledged recruitment departments full of recruiters.
They have set-up processes to
▶ Post Jobs
▶ Search Resumes
▶ Short list
▶ Send out interview-call emails etc. etc.
They can afford all these “processes”
and all these full time “staff”
because
▶ they have many vacancies (New as well as replacements)
▶ vacancies arise around the year with regularity
But SME (Small Medium Enterprises)
▶ have ONE vacancy in a year
▶ no staff or processes
▶ no luxury of “time” to fill vacancy
▶ no budget for jobsite subscriptions
▶ usually want a LOCAL person only
(so, why not eliminate screen for CITY selection?)
By default, Employer’s CITY address to be taken for match-making.
Section
Data Field
Job Search
In Our Website - For Jobseeker Use
Employer Use
RSS Feed
MyJobs Screen
Job Search
Job Alert Summary
Job Alert Detail
Job Post
A
Job Details
✓
✓
▽
✓
✓
Job Title
Design. Level
▽
Designation
✓
✓
✓
✓
Job Description
✓
Category (IT/Non-IT)
◎
◎
Function
▽
▽
Keywords
✓
India Background
✓
▽
✓
Country
✓
▽
✓
City
✓
▽
✓
✓
Currency Offered
✓
Salary Offered
✓
B
MAN-SPES
Candidate Attribute
Work Exp
✓
✓
▽
✓
✓
Edu. Level
✓
✓
Degree / Diploma
✓
✓
✓
✓
Specialization
✓
C
ADVT SETTING
Expiry Date
✓
Job Type
✓
D
EMPLOYER DETAILS
Co. Name
✓
✓
✓
✓
Legend:
✓ = Fully available
▽ = Partially available or in summary form
◎ = Category-based selector or dropdown option
Let me know if you'd like this in tabular format for Word or Excel export.











































